NandhaKishorM/laya

Multilingual, non-autoregressive System 1 decision engine. Typed decisions over 100+ languages in a single forward pass — 33 ms — trained with reinforcement learning against strictly proper scoring rules (RLCD), with a router that picks the right checkpoint per request.

python -m pip install laya

With uv, run uv add laya in a uv project or uv pip install laya in a virtual environment.

Python 3.10 or newer. Optional extras: laya[serve] (HTTP server), laya[mcp] (MCP server), laya[langchain] (LangChain and LangGraph), laya[llamaindex] (LlamaIndex selectors), laya[crewai] (CrewAI routing), laya[onnx] (ONNX Runtime), laya[fast] (TileLang GPU fast path). Step-by-step setup for each platform, CPU-only or GPU PyTorch builds, and troubleshooting are in Installation details.

For TypeScript / Node.js / browser, see laya-ts/. npm releases (npm install laya-ts) are published from this repository's laya-ts-v* release tags.

Long documents. laya-multilingual reads up to 8,192 tokens with max_len=8192. Measured accuracy and time by document length, reproducible with research/scripts/bench_long_context.py:

Long documents: laya-multilingual reads up to 8,192 tokens. It ships with a 1,024-token limit that cuts long documents off, so pass max_len=8192 for them:

result = router.predict(long_document, questions, model="multilingual", max_len=8192)

In the table above, 16 to 18 of 20 requests were answered correctly with up to about 4,000 tokens of text before them; beyond that results vary (8 to 17 of 20), so check long-document accuracy on your own data. Short inputs give identical answers with max_len=8192, and speed follows the input's real length, not the limit: short inputs are unchanged, and a 4,000-token input takes about 1.7 s on an Apple GPU. Name the checkpoint with model="multilingual", since long mostly-English text would otherwise route to the English checkpoint.

from laya import Router

router = Router()  # downloads a checkpoint on first use; Router(preload=True) loads all three up front

state = "Hi, we were billed twice for March. Please refund the duplicate today or we will cancel our plan."
questions = {
    "department": {"type": "choice", "instructions": "Which department should handle this?",
                   "criteria": {"billing": "invoices, payments, refunds",
                                "technical": "bugs, outages, system errors",
                                "other": "everything else"}},
    "urgency": {"type": "score", "instructions": "How urgent is this?",
                "criteria": ["not urgent", "soon", "blocking"]},
    "churn_risk": {"type": "noul", "instructions": "Does the user threaten to cancel or leave?"},
}

result = router.predict(state, questions)
print(result["answers"]["department"]["choice"])  # billing
print(result["answers"]["churn_risk"]["noul"])    # probability the answer is yes
print(result["routing"]["model"])                 # english

The same call works in any of 100+ languages. The Router detects the script and language and sends non-English text to laya-multilingual:

for text in ["मुझसे मार्च में दो बार शुल्क लिया गया, कृपया डुप्लिकेट राशि वापस करें।",
             "La aplicación se cierra cada vez que abro la configuración."]:
    r = router.predict(text, {"department": questions["department"]})
    print(r["routing"]["model"], r["answers"]["department"]["choice"])
# multilingual billing
# multilingual technical

From the command line, laya "My payment failed twice" --preset triage answers a ready-made question set. More in the full quickstart and the docs.

The shipped checkpoints work zero-shot, but fine-tuning on decisions from your own domain is where accuracy jumps. On the typed-decisions benchmark (2,000 decisions across four workflows), the fine-tuned laya-typed-decisions checkpoint scores 0.766 accuracy, against 0.362 for the base English checkpoint on the same decisions.

Fine-tuning notebook (Kaggle 2xT4) and Apple Silicon script (MPS / CPU): run the whole loop (build dataset, train with RLCD, calibrate temperatures, evaluate, export). Details in Fine-Tuning.

nandhakishorm.github.io/laya: guides for prediction hooks, schema-driven decisions, Docker and LangChain and LangGraph, plus a full API reference.

• Concurrency. Router.load builds a checkpoint outside the lifecycle lock, so a cold build no longer blocks calls for a checkpoint that is already resident, and unload is synchronized per checkpoint rather than globally (#848). predict_long holds the tokenizer lock like every other encode, so scanning a document next to ordinary predictions no longer raises Already borrowed (#825).
• Calibration. Per-option-count abstention thresholds, because one min_confidence does not transfer across option counts (#394): fit_abstention_thresholds fits one cut per bucket, keyed like temperature_by_options, and fails closed on a bucket too unreliable to accept anything. Histogram binning (fit_binning_map, apply_binning_map) recalibrates a bucket whose reliability curve temperature scaling cannot reach. save_calibration writes atomically, and records_from_labeled works on a loaded agent (#826).
• Evals. Selective-classification metrics (Brier, AURC, selective accuracy) from the same pairs ECE already uses, so the abstention gate has a yardstick. A NaN metric, limit or tolerance can no longer pass a gate (#830).
• Stricter inputs. A non-finite hooks_timeout (#828), a non-scalar choice label, a batch item's untyped lang_guess, and an unvalidated min_confidence or hooks_timeout on laya_predict_batch are all refused where they are read. hooks_installed accepts a sequence, as documented (#829). A null score level is a 422 instead of a response no Jev client can parse.
• laya-serve. LAYA_JEV_STRICT projects the response onto the strict Jev wire contract for clients that reject unknown fields. /v1/systemone/batch sums output_tokens instead of reporting zero, and refuses an unpaired surrogate like the single endpoint.
• Fallback notes are warnings, not prints. Eight device- and inference-fallback notes went to stdout, which corrupted laya --json output; they are warnings.warn now, so they are filterable.
• One test list. CI and the release gate run the same shared suite list (scripts/test_suites.py). They had drifted to 80 suites against 47, so a regression in any of the 33 CI-only suites could reach PyPI.
• TypeScript. The SDK types and validates the whole usage report, the four confidence fields, and the routing detection's mixed_segment. laya-ts reports collapsed options like laya.common and keeps bytecode out of the npm tarball.
• Docs and examples. Nine example and page corrections where the prose contradicted the code, each with a test that holds the page to the source. The PyPI README links absolutely, so its 31 relative links no longer 404 on the project page.

56 pull requests from 16 contributors.

• Security. GET /health no longer answers deployment internals to an unauthenticated caller on a server that set LAYA_API_KEY (#812): liveness stays open so every shipped probe keeps working, while the resident checkpoint names, revision SHAs, device state and fallback reasons need the bearer. A deployment with no key set is unchanged. SECURITY.md now documents private vulnerability reporting.
• Concurrency. A GPU OOM fallback no longer moves the shared model under another in-flight request (#649). Normal forwards stay concurrent through a reader-writer gate; only the device demotion is exclusive.
• Per-call controls reach every surface. lang, lang_guess, min_confidence, task and the token budgets now forward through the CLI (--lang-guess, --min-confidence), /v1/systemone/batch, the MCP single-request and batch tools, the TypeScript SDK and the LangChain, CrewAI and LlamaIndex wrappers. The three framework wrappers also gate confidence_threshold on answer_confidence rather than the entropy confidence, matching core's own gate.
• Routing. Swedish is detected with MASSIVE evidence across all 51 locales and no locale regressing. LAYA_DEFAULT_MODEL makes the routing fallback settable from the environment, and a per-checkpoint pin no longer silently disables the caller's digest or revision.
• ONNX. --quantize defaults to per-tensor, because per-channel collapsed the decision model to 32 percent agreement with eager. Export declares its dynamic dims in a spelling every torch in the supported range accepts, and verifies the symbolic axes it wrote.
• Long documents. predict_long sizes its windows from the room a question actually leaves for the state, so a scan no longer skips part of the document and reports that it read it.
• Evals. Opt-in per-slice quality gates catch a slice regressing while the overall metric improves, shortlist runs attribute errors to retrieval or decision, --min-confidence reaches the abstention gate, and --calibration works on the ONNX path.
• Abstention reporting. With min_confidence set, every answer reports abstention and abstention_threshold. With no threshold the payload is byte for byte what it was.
• Research. A Spanish phone-turn benchmark (217 frozen sentences, label policy fixed before any model ran) and a Chinese reliability evaluation with source-group isolation and paired bootstrap intervals.
• Email. French mail clients are cleaned the way English, Portuguese and Spanish already were.

64 pull requests from 21 contributors. #742 landed inside #684, which had been rebased onto it. Full list in the 0.3.23 release.

• ONNX catches up with PyTorch. ONNXAgent gains predict_batch (with sort_by_length), predict_long and decide_batch, scripts/export_onnx.py --quantize writes a per-channel INT8 copy for CPU, and laya-evals run --onnx scores an export with the same gates as the torch path.
• Opt-in abstention. min_confidence= on predict, predict_batch, decide and decide_batch flags answers below a threshold on answer_confidence with low_confidence: True, and decide returns None for them. When a threshold is set, every answer also reports abstention — passed, abstained or unevaluated — so a caller can tell a gate that cleared from a gate that never ran; with no threshold, nothing is added at all.
• Batch everywhere. decide_batch, Router.predict_long, laya --batch FILE, the MCP laya_predict_batch / laya_route_batch / laya_decide tools, and LangChain batch() / abatch() all run on shared forward passes. New LayaDecision (LangChain), LlamaIndex selectors (laya[llamaindex]) and CrewAI routing (laya[crewai]).
• Per-request token budget. max_len / head_max_len now reach every surface: laya-serve (capped by LAYA_MAX_TOKEN_BUDGET), Router.predict_batch requests, the CLI (--questions, --max-len, --head-max-len), MCP tools and LangChain nodes.
• Operations. LAYA_MAX_LOADED, LAYA_REVISION and per-checkpoint SHA-256 maps; /health reports the device a checkpoint really runs on and its CPU-fallback count; the 503 busy answer carries Retry-After; compile=True no longer recompiles for every request shape.
• Stricter inputs. A null or duplicate choice label, a short temperature list, a None state and non-dict questions are refused with a message that names them, and usage["options"] says when the head budget left two options with the same tokens.

Laya evaluates typed questions (choice, score, noul) over any state (text, email, ticket or JSON document) in a single forward pass — 33 ms for one question, 7.2 ms/question batched, measured on a T4. No text generation, so nothing to parse and nothing to hallucinate.

Three checkpoints, and a Router that picks between them per request:

Python 3.10 or newer. The dependencies set that floor: huggingface_hub 1.x, transformers 5.x and torch 2.14 all require 3.10.

Optional PyTorch build selection: If you need a CPU-only or GPU-specific PyTorch build, follow PyTorch's installation guide after creating your virtual environment and before installing Laya. Replace pip or pip3 in the selected command with the environment's Python executable followed by -m pip.

If you already use a virtual environment, install the PyPI release with:

python -m pip install laya

For a new environment, choose the commands for your platform below. Run them from your project directory; the explicit Python paths keep installation and verification in the same environment.

macOS / Linux (with Python 3.10 or newer):

On Debian/Ubuntu, the system Python may require sudo apt install python3-venv before creating a virtual environment. If venv reports that ensurepip is unavailable, install that package and retry.

python3 -m venv .venv
.venv/bin/python -m pip install laya
.venv/bin/python -I -c "import laya; print(laya.__version__)"

Windows PowerShell (this example uses an installed Python 3.11):

py -3.11 -m venv .venv
.\.venv\Scripts\python.exe -m pip install laya
.\.venv\Scripts\python.exe -I -c "import laya; print(laya.__version__)"

Both checks print the installed Laya version without loading a checkpoint. -I excludes the current directory from the import search path, so a local source copy cannot mask a missing installation. Keep using the same virtual environment's Python when running your application.

Intel GPU (XPU)

Install a supported Intel GPU driver first. For an XPU-enabled PyTorch build, install its wheel before Laya; the default PyPI wheel may be CPU-only. PyTorch's validated hardware and OS list is in the Intel GPU guide.

.\.venv\Scripts\python.exe -m pip install torch --index-url https://download.pytorch.org/whl/xpu
.\.venv\Scripts\python.exe -m pip install laya
.\.venv\Scripts\python.exe -c "import torch; print(torch.xpu.is_available())"

For a source checkout, replace pip install laya with pip install -e .. Laya automatically selects an available XPU when no device is specified; you can also request one explicitly with device="xpu" in laya.load() or Router(device="xpu").

Install from GitHub

To use the development version instead of the PyPI release, create the virtual environment above and replace its installation command with the appropriate command below. Git must be installed.

# macOS / Linux
.venv/bin/python -m pip install "git+https://github.com/NandhaKishorM/laya.git"

# Windows PowerShell
.\.venv\Scripts\python.exe -m pip install "git+https://github.com/NandhaKishorM/laya.git"

Run the same version check afterward. The GitHub version follows the repository's default branch and may differ from the published release.

Install with uv

uv creates the virtual environment, downloads a matching Python if none is installed, and installs into it. The commands are the same on macOS, Linux and Windows PowerShell:

uv venv --python 3.12
uv pip install laya

uv pip install targets the .venv in the current directory without activating it, so run the version check for your platform above afterward. Extras and the GitHub version install the same way: uv pip install "laya[serve]", uv pip install "git+https://github.com/NandhaKishorM/laya.git". For a CPU-only or GPU-specific PyTorch build, add --torch-backend=auto to pick the build that matches the machine's GPU driver, or name one such as --torch-backend=cpu; uv's PyTorch guide has the list.

If your application is a uv project, add Laya as a dependency instead:

uv add laya
uv run python -I -c "import laya; print(laya.__version__)"

--torch-backend applies to uv pip only; in a uv project, uv's PyTorch guide shows how to set the PyTorch index in pyproject.toml.

Model setup and troubleshooting

Continue with the Router quickstart to run inference. Loading a Hub checkpoint requires access to Hugging Face on its first download; the quickstart's Router(preload=True) loads all three configured checkpoints at construction.

• ModuleNotFoundError: No module named 'laya': run both installation and your script with the same virtual environment's Python executable shown above. In an editor, select that interpreter as well.
• Missing rl_agent_config.json: this file ships with a Laya checkpoint alongside model.safetensors; it is not a configuration file you need to create in the source repository. For a local model, pass the directory containing those checkpoint files.
• Checkpoint downloads and cache: Router() is lazy: the first prediction downloads only the selected checkpoint, then reuses the Hugging Face cache. Set HF_HUB_CACHE to move that cache. The English and typed-decisions checkpoints have about 421M parameters each; multilingual has about 322M (see the table above). Download sizes also depend on weight precision and files.
• Memory limits: Router() keeps up to two checkpoints resident by default (max_loaded=2), evicting the least recently used when needed. Use Router(max_loaded=1) to reduce resident memory, at the cost of reloads when switching checkpoints. Router(preload=True) loads all three up front and raises the resident limit to fit them; it is not an out-of-memory remedy.
• Windows + Python 3.14: the model-construction crash reported in #123 was fixed in Laya 0.3.7 by #195 and verified on the reporter's Windows 11 setup without a manual patch. Upgrade older Laya installations with your environment's Python executable followed by -m pip install -U laya.

Installing the package also installs a laya command for quick local testing, no script needed:

laya "I was charged twice, please refund"            # routing decision only; works offline, no download
laya "Refactor this service" --predict               # full answers (downloads the checkpoint on first use)
laya "Mein Konto wurde zweimal belastet" --lang de   # force a language instead of detecting it
laya "My payment failed twice" --lang-guess en        # a soft hint: nudge routing, still fall through to detection
laya "My payment failed twice" --model ml            # pin a checkpoint: names, aliases and casing all resolve as the SDK resolves them
laya "My payment failed twice" --preset triage       # answer a ready-made preset (triage, email, guard, moderation, router)
laya --batch tickets.txt --predict                   # score a file of requests, one per line, in one batch
cat tickets.txt | laya --batch - --predict --json    # stdin; one JSON line of answers per request
laya "Where is my card" --questions intents.json     # answer your own questions, written in a JSON file
laya "Refund my card" --predict --min-confidence 0.9 # mark an answer the model is unsure of
laya                                                 # interactive mode

--lang is decisive: a real code picks the checkpoint and skips detection. --lang-guess is the soft sibling core checks in between, so a probable-but-uncertain language can nudge routing without forcing it — and a hint that resolves to nothing falls through to the built-in detector, exactly the Router(lang_guess=...) behaviour the library offers. Neither flag changes the routing decision's precedence rules; --lang-guess just adds a hint those rules read before detection.

Routing alone never downloads a checkpoint, so it returns in milliseconds. --predict loads the routed checkpoint, which needs network access to the Hugging Face hub the first time; if a checkpoint cannot be downloaded, the CLI says so instead of crashing. --batch (with or without --predict) sends the whole file through Router.predict_batch in one process, so the requests share checkpoint loads and forward passes — measured 2.6x on 20 tickets vs looping predict one by one, with --batch-size N to bound the forward pass, --sort-by-length to group similarly sized requests inside it, and --json for JSONL output. Batch routing (laya --batch FILE, no --predict) likewise answers with route_batch in one pass, still without loading anything.

--sort-by-length reaches the length grouping Agent.predict_batch has done since #294, which until now only the library call in Batch Mode could ask for. A forward pass pads every state in it to the longest one, so a two-line request sharing a pass with a two-paragraph one spends most of its compute on padding; grouping by length first puts comparable sizes together. Measured on 128 real Yelp reviews of 69-2,293 characters, run as laya --batch FILE --predict --model laya --batch-size 8 with and without the flag, four interleaved rounds on an Apple-silicon Mac at the CLI's own device pick: 41.25 s median unsorted against 29.23 s sorted, a paired median of 1.42x (1.39-1.44x across the four pairs), with all 128 decisions identical and still in input order. research/ measures the same knob at 2.15x over 10,000 synthetic multilingual tickets. It needs --batch-size N with 1 < N < the number of requests — a single pass over the whole file has no group to reorder — and the CLI says so on stderr rather than printing a same-speed result and leaving you to notice.

--questions takes the same question dict the SDK takes, as JSON: either the mapping itself, or {"state_key": "body", "questions": {...}} when the question's instructions name a field other than request. It implies --predict, and a question set with many labels usually wants --head-max-len with it: on 58 MASSIVE-INTENT labels written as one choice question, the English checkpoint goes from 24/58 correct at its default 192-token option budget to 34/58 at --head-max-len 384, for about 1.4x the per-request time on CPU. Widening it further costs the accuracy back, because max_len then leaves fewer tokens for the request itself. Honest limits describes the same budget ceiling for a 77-option question.

examples/server.py is a self-contained FastAPI app for testing Laya without writing any code: a two-pane playground (edit the request as a form or as JSON, run it with Ctrl+Enter, read each answer's full distribution and calibrated confidence, copy it as curl or Python), plus a plain JSON API (/predict, /predict/batch) for scripting against.

pip install "laya[serve]"
python examples/server.py               # http://127.0.0.1:8000

Open http://127.0.0.1:8000 in a browser for the playground, or hit it directly:

curl -s localhost:8000/predict -H 'content-type: application/json' -d '{
  "state": {"body": "We were billed twice for March. Please refund it today."},
  "questions": {
    "department": {"type": "choice",
                   "instructions": "Which department should handle this?",
                   "criteria": {"billing": "invoices, payments, refunds", "other": "everything else"}},
    "urgency": {"type": "score",
                "instructions": "How urgent is this?",
                "criteria": ["not urgent", "soon", "critical"]}
  }
}' | python -m json.tool

--no-preload loads checkpoints lazily instead of all three up front; --device cuda|cpu|mps pins the device. See python examples/server.py --help for the rest.

A request body may carry model to pin a checkpoint instead of letting the router choose, and it takes exactly the spellings laya --model takes: names, aliases and casing all resolve through laya.router. GET /models lists the checkpoints and the aliases alongside them.

Both /predict and /predict/batch also accept the four per-call controls laya-serve forwards on /v1/systemone: lang_guess (a soft ISO hint that participates in routing, where lang skips detection entirely), max_len and head_max_len (per-call token-budget overrides, capped by LAYA_MAX_TOKEN_BUDGET), and min_confidence (core's abstention threshold; the answer comes back with abstention and abstention_threshold fields naming whether it fired). An absent key means "inherit what the Router was built with", so a caller never has to send them back as null; a body that names hooks, hooks_raise, hooks_timeout, on_predict_start, or on_predict_end is refused with 422 rather than silently dropped, matching what laya-serve does for the same fields. On /predict/batch, min_confidence reaches Router.predict_batch as a call argument rather than a per-request key; the other three travel inside each request dict, so states in one batch may name different budgets.

/predict/batch accepts two optional body fields that control the shape of the forward passes without changing any answer: batch_size (states per pass; omit it and the whole batch is one pass) and sort_by_length (group similarly sized states so each pass pads to a shorter maximum — it needs a batch_size below the number of states, since one pass has nothing to reorder). Measured on the running server: 64 real reviews of 91–2,293 characters, one choice question, batch_size: 8, three interleaved rounds on MPS — the default one-pass shape 5.35s median against 3.04s sorted, a paired median of 1.77x (min 1.76x), and 64/64 labels identical in input order. batch_size: 8 on its own was 4.95s, so the length grouping is what earns the second factor: 1.66x over the sized batch.

laya-client supports JavaScript and TypeScript applications over HTTP to a self-hosted laya-serve /v1/systemone server. It includes inferred answer types, all five question presets, ESM/CommonJS exports, and cancellation.

# From this checkout, install and start the Python inference service:
pip install -e '.[serve]'
LAYA_HOST=127.0.0.1 LAYA_MODELS=english laya-serve

# In another terminal:
cd sdk/typescript
npm ci
npm run build
node examples/triage.mjs

The npm package is named laya-client. Use it when a JavaScript or TypeScript application talks over HTTP to self-hosted Python laya-serve. Use laya-ts when inference must run directly inside JavaScript through its local ONNX runtime, without a Python server. See installation, examples, and publishing and the repository analysis for architecture and scope.

To try the Python SDK in a CPU container, see the Docker Compose quickstart. It runs a sample request and keeps downloaded models between runs.

Laya ships three checkpoints. The built-in Router is the recommended entry point: it evaluates any state in any language, automatically detects scripts and languages in sub-milliseconds, and dispatches to the optimal checkpoint in a single forward pass.

from laya import Router

router = Router()  # downloads only the selected checkpoint on first use
result = router.predict(
    {"body": "We were billed twice. Please refund the duplicate."},
    {"billing": {"type": "noul", "instructions": "Does the user request a refund?"}},
)
print(result["answers"]["billing"]["noul"])  # P(true), a float from 0.0 to 1.0
print(result["routing"]["model"])            # which checkpoint answered, e.g. 'english'

Input is any state (str, dict, or list — text, email, ticket, JSON document) plus a dict of typed questions (choice = pick one label, score = ordinal levels, noul = yes/no statement check). Each result["answers"][qid] is an answer dict: choice is the selected label, score is the expected ordinal level, and noul is P(true). confidence is a separate field in that dict; probabilities is available for choice and score. result["routing"] explains which checkpoint was picked and why. Values depend on the checkpoint and input. For ready-made question sets see laya.presets (triage_questions, moderation_questions, email_questions, ...).

Full walkthrough:

from laya import Router

# Preload checkpoints into memory for instant sub-35ms routing
router = Router(preload=True)

# 1. State in any language or schema
state = {
    "from": "user@acme.com",
    "subject": "Duplicate charge on invoice #4411",
    "body": "Hi, we were billed twice for March. Please refund the duplicate today or we will cancel our plan."
}

# 2. Define your typed questions
questions = {
    "department": {
        "type": "choice",
        "instructions": "Which department should handle this request?",
        "criteria": {
            "billing": "invoices, payments, refunds",
            "technical": "bugs, outages, system errors",
            "sales": "pricing, new contracts",
            "other": "everything else"
        }
    },
    "urgency": {
        "type": "score",
        "instructions": "How urgent is this request?",
        "criteria": ["not urgent", "soon", "critical deadline or blocking issue"]
    },
    "churn_risk": {
        "type": "noul",
        "instructions": "Does the user threaten to cancel or leave?"
    },
    "refund_requested": {
        "type": "noul",
        "instructions": "Does the user explicitly request a refund?"
    }
}

# 3. English state -> automatically routed to laya (ModernBERT-large, 39.5 ms)
res_en = router.predict(state, questions)
print("Department :", res_en["answers"]["department"]["choice"])  # -> billing (confidence: 0.94)
print("Routing    :", res_en["routing"]["model"])                 # -> english

# 4. Hindi state -> automatically routed to laya-multilingual (mmBERT-base, 32.8 ms)
res_hi = router.predict({"body": "मुझसे दो बार शुल्क लिया गया, कृपया पैसे वापस करें।"}, questions)
print("Department :", res_hi["answers"]["department"]["choice"])  # -> billing (confidence: 0.86)
print("Routing    :", res_hi["routing"]["model"])                 # -> multilingual

# 5. Explicit override when you want a specific checkpoint
res_td = router.predict(state, questions, model="typed-decisions")

Every result carries full routing metadata explaining why the choice was made:

res_hi["routing"]
# {
#   'model': 'multilingual',
#   'repo': 'convaiinnovations/laya/multilingual',
#   'reason': 'non-Latin script (devanagari, 100% of letters); the English checkpoint cannot read it'
# }

Inspect a routing decision without running any forward pass:

router.route({"body": "Der Kunde wurde zweimal belastet"}, questions).reason
# "Latin script but language looks like 'de', not English"

Very short Latin-script text often carries nothing that identifies its language ("Quero cancelar", "Esqueci minha senha"). Such text goes to default, which is "english" unless you change it. If most of your traffic is not English, set:

router = Router(default="multilingual")
router.route({"body": "Esqueci minha senha"}).model                 # -> multilingual
router.route({"body": "Please refund the duplicate charge"}).model  # -> english

Running one of the shipped servers rather than your own Router, the same setting is LAYA_DEFAULT_MODEL=multilingual — in the environment of laya-serve, the MCP server or the laya-serve container (docs/docker.md). It is the routing fallback only: text the detector can place is routed on what it detects, whatever this is set to.

If a lazy router receives an interleaved workload whose requests route to different checkpoints, calling predict() in a loop can still cause unnecessary checkpoint churn when the required checkpoints exceed the resident cache, for example with max_loaded=1 or when typed-decisions is also used.

Router.predict_batch() routes the full workload first, groups requests by checkpoint, then groups requests with the same question schema within each checkpoint. Each compatible group is dispatched to Agent.predict_batch() so states can share forward passes, and results are restored to the original request order.

requests = [
    {"state": "Please refund invoice 1", "questions": questions},
    {"state": "تم خصم المبلغ مرتين", "questions": questions},
    {"state": "Please refund invoice 2", "questions": questions},
]

results = Router(max_loaded=1).predict_batch(requests)
# results stay in input order while compatible requests are batched by checkpoint

Each item can independently set model, task, lang, lang_guess, or the token budget (max_len, head_max_len). Use route_batch(requests) when you only want the ordered routing decisions without loading any checkpoint. predict_many is an alias for predict_batch.

Requests are validated before model loading. Different requests may use different question schemas; requests sharing a checkpoint, question schema and token budget are passed together to Agent.predict_batch().

Prediction hooks installed on the Router run once per request, as they do for predict(), so a redaction hook rewrites every state before the model sees it. Requests that share a checkpoint run all their start hooks before their shared forward pass; see docs/hooks/lifecycle.md.

You can also bound the Agent-level forward-pass batch size, and forward the length grouping knob to every group (see Batch Mode):

results = router.predict_batch(requests, batch_size=8, sort_by_length=True)

On a shared benchmark (17,416 questions, one T4 GPU, identical questions per model):

The English checkpoint collapses on non-Latin scripts (Khmer scores 0.000 accuracy at 0.952 confidence, the raw-temperature figure; 0.705 as served after the clamp). Because the model stays confident while being wrong, confidence gating cannot save you. Router detects the script in <0.5 ms pure Python before the forward pass.

A cold checkpoint build costs seconds; language detection costs microseconds. The lazy default keeps two checkpoints resident — english and multilingual, the only two automatic routing chooses between — so a language flip costs detection only once each has been built. max_loaded=1 rebuilds the checkpoint it just evicted on every switch (measured at a 7.4 s median reload on CPU and 10.3 s on T4), and traffic that only ever sees one language never builds the second, so the default costs a single-language deployment nothing.

For a server or production app, preload:

# Every checkpoint resident in memory; language flips cost detection only (<1 ms)
router = Router(preload=True)
router = Router(preload=True, device="cuda")

# Or preload only the specific checkpoints you serve:
router.preload(["english", "multilingual"])

# If your app already built an agent, attach it to avoid duplicate VRAM:
router.attach("english", existing_agent)

# Manage resident memory (default keeps two hot: english + multilingual, LRU eviction)
router = Router(max_loaded=3)       # keep all three hot, e.g. with auto_task_detection
router = Router(max_loaded=1)       # memory-constrained host, reloads on every switch
router.unload()                     # free memory

A rebuild still re-reads the checkpoint, but each checkpoint's tokenizer is parsed once per process and reused by every Agent — including one the Router rebuilds after eviction. The multilingual tokenizer.json alone is 34 MB / 256k vocab, several times the cost of applying its weights. Preloading is still the right answer for a server: it removes the rebuild rather than making it cheaper.

Routing asks one question: can the English checkpoint read this state? The built-in detector answers it from the script and a function-word heuristic, and is deliberately dependency-free. That heuristic is best-effort on Latin-script languages it holds no word list for, so a short request can carry no usable signal:

from laya.lang import analyse
analyse("Care este ora in Tokyo?")
# {'script': 'latin', 'language': 'en', 'is_english': True}   -> the English checkpoint

If you already run a language-identification model, hand routing the answer instead of relying on the heuristic. lang_guess takes a language code or a callable receiving the state, and is checked after an explicit lang= and before detection:

# A code you already know
router.predict(state, questions, lang_guess="ro")

# A callable, e.g. wrapping fastText, CLD3 or a transformer LID
router.predict(state, questions, lang_guess=lambda s: my_lid(s))

# Or install one for every request on a server
router = Router(preload=True, lang_guess=my_lid)

The hint only decides English or not: a code whose primary subtag is en, eng or english routes to the English checkpoint, and every other code that names a language routes to the multilingual one. "en_US" and "en_US.UTF-8" are read as English, so $LANG can be passed straight through. Returning None, or a code that names no language, makes it abstain and the built-in detector decides as before — so a LID model that is unsure does not force a checkpoint. C, POSIX and C.UTF-8 abstain, which matters because C.UTF-8 is the default $LANG in the official Python image: passing it through no longer pins every request to the multilingual checkpoint, which is what it used to do. The ISO 639-2 special codes und, zxx and mul abstain for the same reason. An explicit model=, task= or lang= still wins, and the default path is unchanged.

laya.serve exposes the Router over HTTP on the same POST /v1/systemone wire protocol as TypeSafe's hosted Jev API. Laya's answer payload is already schema-identical to what Jev returns (choice/score/noul answers and a {input_tokens, output_tokens} usage block), so an existing Jev client — e.g. the hs-jev Haskell client — just needs its baseUrl repointed; nothing else changes. A PHP client that targets laya-serve rather than Jev also exists: marcreichel/laya-php.

pip install "laya[serve]"          # adds fastapi + uvicorn + python-multipart
LAYA_DEVICE=cuda LAYA_PRELOAD=1 laya-serve   # binds 0.0.0.0:8000, preloads all 3 checkpoints

curl -s localhost:8000/v1/systemone -H 'content-type: application/json' -d '{
  "state": {"body": "billed twice, refund please or we cancel"},
  "questions": {"dept": {"type": "choice", "instructions": "which team?",
                "criteria": {"billing": "refunds", "tech": "bugs"}}}
}'

To evaluate multiple states in a single call, send a states array to POST /v1/systemone/batch (capped at 64 states):

curl -s localhost:8000/v1/systemone/batch -H 'content-type: application/json' -d '{
  "states": [
    {"body": "billed twice, refund please"},
    {"body": "cannot login, getting 500 error"}
  ],
  "questions": {"dept": {"type": "choice", "instructions": "which team?",
                "criteria": {"billing": "refunds", "tech": "bugs"}}}
}'

The endpoint shares forward passes via Router.predict_batch and returns an array of results in matching order along with aggregated total_usage.

Configuration is by environment variable: LAYA_HOST, LAYA_PORT, LAYA_DEVICE, LAYA_PRELOAD, LAYA_MODELS (comma list to preload), LAYA_THREADS (cap torch intra-op threads for CPU inference — keep at or below physical cores), LAYA_AUTO_TASK, LAYA_DEFAULT_MODEL (the checkpoint a state with no language evidence falls back to, english by default; set it to multilingual when most of your traffic is not English), LAYA_MAX_LOADED (checkpoints resident at once, 2 by default; raise it to 3 when LAYA_AUTO_TASK makes a third one reachable on demand, or the server rebuilds one every time routing switches), and LAYA_API_KEY (when set, clients must send Authorization: Bearer <key>). A client's model field is honoured when it names a Laya checkpoint (english/multilingual/typed-decisions), otherwise the router auto-selects by script/language. The same body can also carry task, lang, lang_guess, min_confidence, max_len and head_max_len — the controls Router.predict takes that a JSON body can state — while the five hook arguments (hooks, on_predict_start, on_predict_end, hooks_raise, hooks_timeout) are refused with 422, because a hook is a callable that has to live where the server runs.

When publishing the server under a reverse-proxy prefix such as /laya, set LAYA_ROOT_PATH=/laya. The proxy should strip that prefix before forwarding; the setting controls FastAPI-generated URLs and leaves the internal routes unchanged.

Three things differ from Jev when you port a client:

• Options per question. A question's options share the checkpoint's option budget, head_max_len (192 tokens on laya, 256 on the other two), not Jev's cap of 255 options. In addition, the HTTP server (laya.serve) enforces an amplification guard of at most 100 choice options per question (MAX_CHOICE_OPTIONS = 100, rejected with 413 before inference). Once options overflow the token budget, around 20 options with a short description each, every option is trimmed to fit, so long or similar labels can reach the model reading the same (Where Jev leads). Once they no longer fit the window at all, the library rejects the request with 422. For more candidates, narrow them first with predict_shortlist (Honest limits).
• Score levels. Every level needs a description. A null level is rejected with 422 rather than scored and echoed back in legend.
• confidence on choice and score answers is 1 minus normalised entropy, a measure of how concentrated the distribution is, not Jev's (n·p_max − 1)/(n − 1). A threshold carried over from Jev does not transfer. For one calibrated number on every question type, gate on answer_confidence, the probability of the reported answer. Both values shift with the number of options, so fit and validate any cutoff on held-out data at the option counts your workload uses (#394).

This repo is a flake. On a machine with an NVIDIA GPU:

nix run .#laya-serve          # build (prebuilt CUDA torch, no compile) and serve
nix develop                   # dev shell: torch-bin, transformers, fastapi, pytest

For a NixOS host, import the module and enable the service:

# flake inputs:  laya.url = "github:<you>/laya";  # or path:/… on the same host
{
  imports = [ laya.nixosModules.default ];
  services.laya-serve = {
    enable = true;
    host = "0.0.0.0";           # or bind to the Tailscale/LAN address
    openFirewall = true;
    device = "cuda";
    models = [ "english" "multilingual" "typed-decisions" ];
    # apiKeyFile = config.age.secrets.laya-api-key.path;  # optional bearer auth
  };
}

The module runs a hardened DynamicUser systemd unit with CUDA device access, caches weights under /var/lib/laya-serve, and reads the bearer token (if any) via LoadCredential so it never enters the store.

If you only need a single checkpoint for a dedicated pipeline, you can load models directly:

import laya

# 1. Load a specific checkpoint directly from the hub
agent = laya.load("convaiinnovations/laya")                           # English root
agent_ml = laya.load("convaiinnovations/laya", subfolder="multilingual") # 100+ languages
agent_td = laya.load("convaiinnovations/laya", subfolder="typed-decisions")

# 2. Run all questions in ONE single forward pass (~35 ms on GPU)
result = agent.predict(state, questions)
answers = result["answers"]

print("Department :", answers["department"]["choice"])   # -> billing (confidence: 0.94)
print("Urgency    :", answers["urgency"]["score"])        # -> 1.84 / 2.0
print("Churn Risk :", answers["churn_risk"]["noul"])       # -> 0.892 (89.2% probability)

Passing an empty question dictionary to agent.predict(state, {}) or agent.system_one(state, {}) returns the standard response with "answers": {} and "usage": {"input_tokens": 0, "output_tokens": 0}. The state is not tokenized and no model forward pass runs.

predict handles one state per call, which leaves most of the GPU's batch dimension idle. When you have a list of items to score against the same questions — a backlog of tickets, a table of rows, a log slice — predict_batch packs them into shared forward passes:

states = [{"body": t} for t in ticket_texts]           # a list of states

results = agent.predict_batch(states, questions)       # one forward pass for the whole list
# results[i] corresponds to states[i], with the same output shape as predict

# Bound peak memory when the list (or the texts) are large — chunk into passes of N:
results = agent.predict_batch(states, questions, batch_size=64)

# Reduce padding when input lengths vary; results still follow the original state order:
results = agent.predict_batch(states, questions, batch_size=64, sort_by_length=True)

sort_by_length=True groups states by their longest encoded question row, after truncation. It looks ahead at most eight batches and reuses the encoded rows for sorting. This uses more temporary CPU memory for tokenized inputs, and takes effect only when 1 < batch_size < len(states). Benchmark it on your workload and backend: uniform lengths offer little benefit, and changed batch shapes can cause small floating-point differences, including near decision thresholds. Hooks still see states and final results in input order. The option is available on Agent and ONNXAgent.

Results are aligned with states by index and identical in shape to predict. Changing batch shapes can introduce floating-point differences on CPU and GPU; check decision thresholds on your workload, particularly with mixed precision. Batching is a GPU throughput win — on an RTX 5060 Ti, per-decision latency drops from ~10 ms one-by-one to ~1 ms batched (measured ~9–10×). On CPU, increasing batch size alone may not speed up inference; length grouping can help by reducing the padded work in a mixed-length workload. See the CPU measurements and reproduction commands.

ONNXAgent.predict_batch(states, questions, batch_size=..., sort_by_length=...) has the same contract, backed by one ONNX Runtime session run per chunk, so an ONNX deployment gets the same batch API, the same result shape, and the same length grouping. Measured on the English checkpoint (flat fp32 export, Apple M1, 80 support tickets alternating short and ~8x-longer documents, batch_size=4, best of 3): wall clock 53.7 s unsorted → 36.3 s sorted (~1.48x), with 0/80 decision changes and max probability drift 0.0 across queue/noul/score.

predict/system_one truncate a state that exceeds max_len to a single window (the first, or for a conversation list the last), silently dropping the rest. predict_long scans the whole state in overlapping windows, scores them in shared forward passes (via predict_batch), and aggregates per question:

result = agent.predict_long(state, questions)              # windows the state, one result back
result = agent.predict_long(state, questions, window=256)  # smaller window isolates a localized span
result = agent.predict_long(state, questions, hooks=[AuditLog()])   # the scan, instrumented

• noul takes the strongest window (the statement holds if any window supports it).
• choice / score take the most-confident window — averaging over a long, mostly-neutral document lets the neutral majority out-vote the one window that saw the deciding span.
• A state that already fits one window is passed straight to system_one (identical output, plus usage["windows"] = 1). The key is total: 1 single window, N scanned windows, 0 a hook answered the document before the model read any of it.
• Hooks wrap the inference that answers the state, so on a scanned document on_predict_start fires once with ctx.states holding the decoded windows, not the state you passed in (it was tokenized to produce them). A start hook may replace that list: the answers are aggregated over whatever reached inference, and usage["windows"] counts those states. What a rewritten scan costs is the attribution — answer["window"] names a span of your document, so it is only reported when the scan reached inference unchanged. A hook that means to answer the document calls ctx.skip(...) instead: its result comes back with no answer["window"] and usage["windows"] at 0, because no window scored it.

A smaller window isolates a short deciding span better (it becomes a larger fraction of its window); the default (max_len - head_max_len) favors context and throughput. Either way the window is capped at the room the questions leave for the state inside max_len, so a window is never re-truncated on the way to the model and token_start/token_end describe the span it read: many options leave little room (on the English checkpoint, 2 options leave 483 state tokens and 100 leave 100), and a window past that room used to be cut short silently. Output shape matches predict, with usage["windows"] added.

ONNXAgent.predict_long(state, questions, window=..., stride=..., batch_size=...) has the same contract and the same aggregation rules, with the windows scored through ONNXAgent.predict_batch — one ONNX Runtime session run for all of them, or one per chunk when batch_size bounds memory.

The returned probability is the deciding window's, not a calibrated number for the whole document — a noul max drifts up with the window count even with no signal, and choice can land on a confidently-neutral window when nothing is decisive. Each answer carries answer["window"] (the deciding window's index, token_start/token_end, and count) so you can check the span the answer actually came from:

r = agent.predict_long(state, questions)
r["answers"]["refund"]["window"]   # {'index': 13, 'token_start': 4680, 'token_end': 5432, 'count': 14}

The same scan is reachable from the Router, which routes first and then windows the checkpoint it picked — the same model=/task=/lang= hints, hooks and routing key as predict:

result = router.predict_long(state, questions, model="multilingual")

pip install laya[fast] adds an optional forward built from fused TileLang kernels: GEMM + bias/activation epilogues, GEMM + GEGLU, residual + LayerNorm, in-place RoPE, and a sliding-window flash attention that reads the packed QKV buffer directly. Weights stay resident in bf16 and every (batch, length) bucket is captured as a CUDA graph, so a one-question call no longer pays ~200 kernel launches from Python.

agent = laya.load("convaiinnovations/laya", fast=True)   # or: agent.accelerate()
agent.predict(state, questions)                            # same API, same answers

Numerics: on a fixed set of 60 states the fast path stays within 0.046 of an fp32 forward and within 0.076 of the stock bf16 path (max |Δp| ≤ 0.05 vs fp32 on both checkpoints, argmax agreement ≥ 47/48 per question type; every per-option probability is in benchmarks/results/parity_*.json) — see benchmarks/parity_fast.py and BENCHMARKS.md. The fast path runs in the agent's autocast dtype at the time accelerate() is called: bf16 by default, fp16 if agent.dtype is torch.float16, where it stays within 0.009 of fp32 on the same set (BENCHMARKS.md). Falls back to the stock forward on CPU/MPS or when tilelang is not installed; agent.deaccelerate() restores it. Kernels compile once per shape bucket on first use (a few seconds, cached on disk).

compile=True compiles on the first request that needs a graph, and the first single-question request needs a second one (torch specialises a batch of 1). Both stall a live request. agent.warmup() runs the forward on a few synthetic shapes now and returns the seconds it took, so the compiles happen before traffic arrives:

agent = laya.load("convaiinnovations/laya", compile=True)
agent.warmup()                   # ~46 s on an RTX 4070 Ti SUPER; every later request ~10-30 ms

Measured with benchmarks/bench_compile.py --device cuda [--warmup] (English checkpoint, torch 2.11, ten requests of changing shape): without it the first request took 51 s and the first single-question request, the eighth, took another 41 s; after warmup() no request took more than 30 ms. It works the same with fast=True (kernels and CUDA graphs for those buckets) and costs a few forward passes on the stock path. On CUDA, compiled inference pads the masked end of each sequence to a multiple of eight tokens. This avoids extra SDPA graph specialisations on PyTorch versions that distinguish lengths modulo eight; reported token usage still counts the original, unpadded sequence. Inductor caches compiled graphs under TORCHINDUCTOR_CACHE_DIR (by default in /tmp); point it at a persistent directory to keep them across restarts.

Because Laya's probabilities are trained with strictly proper scoring rules (RLCD), confidence scores are statistically meaningful:

dept = answers["department"]["choice"]
conf = answers["department"]["confidence"]

if conf >= THRESHOLD:               # refit and validate THRESHOLD on your own held-out data
    route_automatically(dept)       # above it: act, and sample the decisions you act on
else:
    escalate_to_human_agent(dept, reason=f"Low confidence ({conf:.2f})")

A threshold is a policy you choose from measured accuracy at that coverage on your data, not a property of the model. Both checkpoints are over-confident as shipped and laya-multilingual has no fitted temperatures at all, so fit them before relying on these numbers — see Calibration below, and the fine-tuning notebook for the fitting loop itself. Then pick the point where the errors you accept are ones you can live with. Confidence orders decisions; it does not establish that a decision is correct.

A threshold also depends on the autocast dtype. On CUDA at compute capability 8 or above the runtime uses the checkpoint's amp_dtype, which is bf16 for all three shipped checkpoints. On the fixed set from benchmarks/parity_fast.py (60 states, 288 questions per checkpoint, RTX 2000 Ada) bf16 moves a probability by up to 0.073 against the fp32 forward and flips 3 of 864 argmaxes across the three checkpoints; fp16 stays within 0.019 and flips none, at the same latency. LAYA_CUDA_AMP=fp16 selects fp16 and LAYA_CUDA_AMP=bf16 selects bf16 (LAYA_CPU_AMP=bf16 is the CPU counterpart). MPS autocasts in fp16 too, but the overhead dominates on a single small row, so there it engages only once a call reaches mps_amp_min_rows rows -- 5 by default, LAYA_MPS_AMP_MIN_ROWS to move it; a value that does not parse falls back to 5 and anything below 1 is clamped to 1. On MPS, agent.dtype is the autocast target, so it says float16 even when every call stays below the gate and runs in fp32; agent.dtype_for(rows) returns the precision a call with that many rows runs in. Fit and measure a threshold in the dtype you serve with.

predict, predict_batch, system_one and decide — on Agent, Router and ONNXAgent — take an opt-in min_confidence, off by default. It is a caller-side policy on top of the emitted confidence: every answer whose answer_confidence falls below the threshold is flagged low_confidence: True, with the raw answer, probabilities and confidence left intact for inspection.

res = agent.predict(state, questions, min_confidence=0.85)
ans = res["answers"]["department"]

if ans.get("low_confidence"):        # answer_confidence < 0.85
    escalate_to_human_agent(ans["choice"], reason=f"Low confidence ({ans['answer_confidence']:.2f})")
else:
    route_automatically(ans["choice"])

The threshold reads answer_confidence (max(p)) — the calibrated quantity, invariant to the number of options — never the entropy confidence. With decide(..., min_confidence=...) a low-confidence field comes back as None in the schema output, while return_details=True keeps the answer and its confidence. LangChain LayaRouter's confidence_threshold reads the same value: answer_confidence when the answer carries it, confidence otherwise. Left unset, min_confidence changes nothing.

A gate is a policy, and a policy whose application you cannot observe is not one. low_confidence is written only when the gate fires, so its absence cannot tell you "a gate ran and this answer cleared it" apart from "no gate ran at all" — you could neither compute an abstention rate nor prove the gate was in effect. When you pass min_confidence, every answer therefore also reports abstention:

abstention_threshold echoes the threshold whenever a gate ran, so a log can be re-split by the gate that produced it instead of by whatever the caller happened to remember passing. unevaluated is the case a boolean cannot express — reporting it as a pass would be as wrong as reporting it as a flag, so a NaN, a missing confidence or a bool lands in its own state. The raw answer, probabilities and confidence are untouched either way.

Left without min_confidence, the response is byte-for-byte what it was before any of this: no abstention, no abstention_threshold, no flag. That is deliberate — a field on every answer that only some calls populate is a schema change for callers who never asked to be gated. The presence of abstention is how you tell a gated run from an ungated one; a min_confidence of exactly 0.0 still counts as set, so it reports states, and flag_low_confidence treats 0.0 as a no-op because nothing can fall below it.

The same gate is a flag on the laya command:

laya "Is it the blue one or the green one" --preset triage --min-confidence 0.9

intent      : other (p=0.992)
frustration : 1.06  [low-confidence]
churn_risk  : 0.158  [low-confidence]

Each marked answer gets [low-confidence] on its printed line and the answer itself is still printed, so a run at a terminal reads the same way the dict does; --json carries the raw low_confidence and abstention keys. --batch FILE applies one threshold to the whole file, which is the shape predict_batch takes. Routing has no answer to gate, so laya --min-confidence 0.9 "..." without --predict, --preset or --questions is refused rather than ignored.

Hooks observe or shape every decision without forking: audit logging, PII redaction before inference, caching, metrics, confidence gating, routing overrides, and forwarding to an external service. They are opt-in, and unset hooks are a no-op.

import laya

def log(ctx):
    print(ctx.model, ctx.results[0]["answers"], ctx.elapsed_ms)

agent = laya.load("convaiinnovations/laya", on_predict_end=log)
agent.system_one("I was charged twice.", {"urgent": {"type": "noul", "instructions": "Urgent?"}})

A hook is a plain callable, or an object implementing any of on_predict_start, on_predict_end, on_route, on_load, on_evict, on_error. A start hook can rewrite the state/questions or ctx.skip(...) a cached answer; an end hook can rewrite the result. See docs/hooks/ and examples/hooks/.

Describe the shape you want with a JSON schema or a pydantic model, and Laya answers it in one forward pass, with typed values and calibrated confidence.

import laya

schema = {
    "type": "object",
    "properties": {
        "department": {"type": "string", "enum": ["billing", "support", "sales"],
                       "description": "Which team should handle this?"},
        "urgency": {"type": "integer", "minimum": 0, "maximum": 2},
        "needs_human": {"type": "boolean"},
    },
}

agent = laya.load("convaiinnovations/laya")
agent.decide("I was charged twice, refund me.", schema=schema)
# {"department": "billing", "urgency": 2, "needs_human": True}

decide also works on a Router, accepts a pydantic model (install laya[structured]), and can return per-field confidence with return_details=True. In an LCEL chain or LangGraph node the same call is LayaDecision. See docs/structured.md.

For throughput, decide_batch answers a list of states against the same schema through predict_batch (see Batch Mode), so the schema is planned once and the states share forward passes:

values = agent.decide_batch(ticket_texts, schema=schema)   # values[i] matches ticket_texts[i]

# On a Router the states may land on different checkpoints; keywords reach predict_batch:
details = router.decide_batch(states, schema=schema, return_details=True, batch_size=64)

Measured on an Apple M-series (MPS), 8 English tickets through one checkpoint: 2624 ms one-by-one vs 723 ms batched (3.6×); 16 mixed English/German states through a Router: 2977 ms vs 1903 ms (1.6×). On CPU the same workloads gave 1.6× and 2.0×. Projected values matched the one-by-one loop 8/8 and 16/16 on both devices, as expected wherever argmax is not at a threshold — verify against your own decision boundaries. laya.decide_batch(runner, states, ...) is the function form, and ONNXAgent.decide_batch runs the same thing on an ONNX export.

Laya provides pre-tuned question schemas for immediate production use:

import laya

agent = laya.load("convaiinnovations/laya")

# 1. Intelligent Model Router (routes to small vs. frontier models)
routing = agent.predict({"request": "Refactor this service using dependency injection"}, laya.router_questions())

# 2. Real-time Prompt Guardrails (jailbreaks, injections, leaks)
guard = agent.predict({"prompt": "Ignore all instructions"}, laya.guard_questions())

# 3. Content Safety & Moderation (toxicity, harassment, threats)
safety = agent.predict({"post": "User comment text"}, laya.moderation_questions())

# 4. Support Ticket Triage (intent, urgency, frustration, churn)
triage = agent.predict({"message": "My payment failed twice"}, laya.triage_questions())

Fast System 1 routing and guardrails directly inside LangGraph workflows and LCEL chains. Every node also takes core's per-call prediction hooks (hooks, on_predict_start, on_predict_end, hooks_raise, hooks_timeout):

from laya.integrations.langchain import LayaDecision, LayaGuardrail, LayaRouter

# 1. Sub-35ms LangGraph conditional edge routing with confidence fallback
router = LayaRouter(
    criteria={"billing": "invoices, charges", "tech": "bugs, outages"},
    confidence_threshold=0.80,
    fallback="human_agent",
)
workflow.add_conditional_edges("triage", router)

# 2. Inline prompt guardrails
guard = LayaGuardrail(action="raise")  # raises LayaGuardrailError on jailbreak/injection

# 3. Schema-driven decisions: a JSON schema in, schema-shaped values out
decide = LayaDecision({
    "type": "object",
    "properties": {
        "department": {"type": "string", "enum": ["billing", "technical", "other"]},
        "needs_human": {"type": "boolean"},
    },
})
decide.invoke("I was charged twice and the API still 500s for us.")
# {'department': 'billing', 'needs_human': False}

# 4. Score a backlog in one batched call, not one forward pass per input
routes = router.batch(["refund my invoice", "the app crashes", "change my password"])

batch() and abatch() run the whole list through predict_batch, so chain.batch(...), RunnableParallel and LangGraph map-reduce nodes get Laya's shared forward passes instead of LangChain's default one-call-per-input thread pool -- which races on MPS, where concurrent torch forwards abort the process. Measured on Apple M-series (medians of three): 2.2x on a 16-ticket routing batch, 2.2x across 24 mixed-language tickets through a Router, 1.8x on the guard preset, with every route label and guardrail flag unchanged. On CPU the same workloads are 2.2-2.4x over the one-by-one loop and 1.1-1.5x over the thread pool.

See docs/langchain.md for the full guide, support ticket triage nodes, and remote HTTP server configuration.

noul always scores two semantic slots in [false, true] order and returns the probability of the second slot. For compatibility, those slots are shown to the model as false and true by default. The optional labels mapping overrides only that model-facing text without changing the returned meaning:

question = {
    "type": "noul",
    "instructions": "Is this review positive?",
    "criteria": {
        "false": "the review is negative",
        "true": "the review is positive",
    },
    "labels": {
        "false": "B",
        "true": "A",
    },
}

The labels mapping is optional. It must contain exactly the string keys false and true, whose values must be distinct non-empty strings. Mapping order does not matter, and the returned noul value is still P(true). Label sensitivity varies by checkpoint and state, so validate any override on your own data rather than treating A/B as a universal fix.

Every question type also takes an optional option_order: a permutation of the option indices saying which option goes in which slot. Slot s shows option option_order[s]. Probabilities always come back keyed in your own option order, whatever order the model saw them in, so the key is presentation only — it never changes what a returned label means.

question = {
    "type": "choice",
    "instructions": "Which team should handle this?",
    "criteria": {"billing": "...", "technical": "...", "sales": "..."},
    "option_order": [2, 0, 1],     # slot 0 shows `sales`, slot 1 `billing`, slot 2 `technical`
}

It exists because where an option sits changes the answer. research/eval/presentation_checks.py measures the English checkpoint on five options whose text is identical: the per-slot logits centre at [+1.68, +1.47, +0.11, -1.35, -1.91], a spread decided by position alone, and the multilingual checkpoint leans the other way on score (see #131). The checkpoint-side fix is a position-balanced retrain; until then option_order lets a caller average the effect out, by asking the same question under orders in which every option occupies every slot equally often:

k = len(question["criteria"])

# the k rotations go in as k questions, so they share one forward pass
rotations = {
    "q%d" % r: dict(question, option_order=[(i + r) % k for i in range(k)])
    for r in range(k)
}
answers = agent.predict(state, rotations)["answers"]

totals = {label: 0.0 for label in question["criteria"]}
for answer in answers.values():
    for label, value in answer["probabilities"].items():
        totals[label] += value / k

decision = max(totals, key=totals.get)

predict already scores every question in a request in one parallel forward pass, so this costs k rows rather than k forward passes — sending the rotations as separate predict calls would cost k of them. On 62 banking intents shortlisted to 20 over 49 states, averaging this way took the share of answers that change with presentation order from 16.3% to 6.1%. Omitting option_order keeps the canonical order and the behaviour every existing caller already has.

Laya can be exposed as an MCP stdio server, so any MCP client (OpenClaw, Claude Desktop, Cursor, ...) can call typed decisions as tools (laya_predict, laya_predict_batch, laya_route, laya_route_batch, laya_decide, laya_shortlist, laya_preset, laya_status) without writing glue code. This is an optional extra: the core package has no mcp dependency.

pip install "laya[mcp]"
laya-mcp-server          # or: python -m laya.mcp.server

Example MCP client configuration (stdio transport):

{
  "mcpServers": {
    "laya": {
      "command": "laya-mcp-server",
      "env": { "LAYA_DEVICE": "cpu" }
    }
  }
}

The environment variables follow the contract documented at the top of laya/serve.py, so the same variable has one meaning across the package:

The tools return structured JSON (answers with probabilities, routing metadata, device, latency_ms). laya_predict_batch and laya_route_batch are the MCP form of Router.predict_batch / route_batch: one tool call takes an array of {state, questions, model?, task?, lang?, lang_guess?, max_len?, head_max_len?} requests, routes them first, groups them by checkpoint, and shares forward passes between requests with the same question schema, returning the answers in input order. max_len / head_max_len are per request there, exactly as Router.predict_batch reads them, so one wide question can raise its own budget without shrinking the rest of the batch to it. On 16 mixed-language tickets through the tool functions themselves, one batch call beat 16 laya_predict calls by 2.1-2.3x on MPS (983-1082 ms -> 467-477 ms) and ~1.25x on CPU (1861-2471 ms -> 1470-1911 ms), with 0/16 decision flips (choice label, rounded score, noul sign) against the loop. Prefer it whenever a client has more than a few requests: each saved round trip is also an MCP request/response. laya_decide is the MCP form of laya.decide: it takes a JSON schema (enum choices, booleans, bounded integers) instead of hand-written questions and returns the decided values projected onto that schema -- enum member, integer level, boolean -- beside per-field confidence and probabilities, so a client that already knows the answer shape never parses an answer map by hand. laya_shortlist is the MCP form of predict_shortlist: it shortlists a many-option choice question to its k most likely labels by embedding similarity (mean-pooled from the answering checkpoint's own encoder, so no extra model is downloaded), answers in one forward pass, and returns per-question shortlist metadata (kept labels, cosine scores, k, option count). The guardrails shown on every decision tool point clients to laya_shortlist for >20-option choices. As with the SDK, use it for structured decisions only; not for open Q&A or text generation. Tests: tests/test_mcp.py (CI, no weights) and tests/test_mcp_local_e2e.py (local, real weights and a real stdio handshake).

laya.evals scores a labelled dataset and gates a build on it, so a quality change is a reviewable diff instead of a hand-check. It is pure Python plus numpy, imports no torch, and needs no weights until you point it at a checkpoint.

laya-evals validate research/evals/fixture.jsonl
laya-evals run data.jsonl --model english --min-accuracy 0.8 --max-ece 0.05 --slice language
laya-evals run data.jsonl --baseline baseline.json --tolerance choice_accuracy=0.02 --json report.json

run reports overall and per-slice metrics (choice_accuracy, noul_accuracy, score_mae, ece, mean_confidence, latency) and exits non-zero on a threshold or baseline failure, so it drops into CI unchanged. A weight-free job runs the metric and API tests on every PR, and a scheduled workflow evaluates the English checkpoint against the committed baseline. See docs/evals.md for the dataset format and the gate.

Community diagnostics: Chinese workplace decisions (Feishu-style) · 中文说明. Includes frozen synthetic cases, archived paired Laya/Jev responses, and an offline audit; separate from the benchmark suites below. Also Chinese short-command routing · 中文说明: 18 frozen commands and a seven-rung ablation of the documented prompt guidance, which locates the accuracy loss on the four-question path rather than the six-option one.

Full report: BENCHMARKS.md — every run consolidated, languages and themes, with per-language detail for all 51 languages.

All Laya numbers below are measured. Every model answered byte-identical questions (fixed seed) in the same run. Reproduce with research/scripts/laya_benchmark_colab.ipynb on a T4.

Batched throughput reaches 103-332 questions/sec on a single T4. For reference, TypeSafe Jev has been independently measured at 236-276 ms p50 (AbdelStark, nibzard) -- Laya answers a single question roughly 6-7x faster.

Every Laya figure is what Router().predict(...) actually returns — the checkpoint the router selects for that input, not a hand-picked best of three. Jev figures are third-party published, never measured here (no TypeSafe API access), so sample sizes and prompts differ. The Laya column here comes from the Applications run (research/scripts/bench_apps.py, N=400 per task; research/results/app_benchmark_results.json), so AG News / DAIR Emotion read 0.950 / 0.595 here versus 0.947 / 0.573 in the committed T4 English suites (N=600) below.

On DAIR Emotion, Jev assigned zero probability to the true label on 16% of examples — a hard failure for anything branching on confidence.

• High-cardinality label spaces (>20 options at default settings): On Banking77, Jev scores 0.870 (on 72 labels) while Laya scores 0.425 (on 77 labels at default 256-token head budget). This is an architectural token-budget constraint: options share a fixed head_max_len budget (192 tokens on English, 256 on multilingual), so 77 options receive only ~3 to 4 tokens per label, causing text to become indistinguishable. Jev supports up to 255 options out-of-the-box. While laya-multilingual supports 1,024 context (and up to 8,192 in the encoder) and you can raise agent.cfg["head_max_len"] = 512 at runtime, Jev is currently better suited for 50+ options in a single prompt without tuning. predict_shortlist (see Honest limits) keeps the top k labels with a caller-supplied embedding, then runs one forward pass on that shortlist.
• Soft distribution matching: On typed-decisions, while Laya achieves higher argmax accuracy (0.766 vs 0.727), Jev achieves higher soft accuracy (0.580 vs 0.471) against the teacher's full probability distributions.
• Out-of-the-box raw calibration: Before temperature scaling, the base checkpoint has higher raw ECE (0.213 vs 0.144). Laya achieves its 0.081 ECE after domain temperature fitting.

Full detail, including every workflow and all 51 languages: BENCHMARKS.md.

400 cases, 2,000 decisions, four workflows.

The fine-tuned checkpoint beats Jev by 3.9 points and clears the teacher ceiling, with 2.4x better Brier and 1.6x better score MAE. It wins on all four workflows: invoice processing 0.804, security incidents 0.766, customer service 0.764, agent-trace observability 0.730. By primitive: noul 0.857, choice 0.733, score 0.723.

Two places it still trails Jev: soft accuracy (0.471 vs 0.580 — its argmax is better but its distributions match the teacher less well) and ECE (0.213 vs 0.144), which temperature fitting addresses.

The base checkpoints sit below the majority-class baseline (0.362 and 0.352 against 0.461). All of the capability on this benchmark comes from fine-tuning. Base rows match the committed runs (suites.typed_decisions in research/results/t4_colab_benchmark.json: 0.3620 / 0.3515; CPU sweep part_b: 0.3615 for English). The laya-typed-decisions row and its four per-workflow scores are from the fine-tuning run and have no committed result file behind them yet.

Across all 51 languages the English checkpoint macro-averages 0.227 with macro ECE 0.733 (0.571 as served, after the temperature clamp), and only 23 of 51 languages clear 3x random. laya-multilingual reaches 0.401 on the same 5,100 cases, clearing 3x random in 48 of 51 — see BENCHMARKS.md for why those multilingual numbers are a re-run rather than the committed file. Khmer scores 0.000 at 95.2% confidence raw, 70.5% as served, on the English checkpoint. This is why Router exists: the model's own confidence gives no warning, so the routing decision has to be made before the forward pass.

Both checkpoints are over-confident as shipped. Refitting one temperature per (question type, option count) on held-out data moves mean ECE 0.466 -> 0.081 (laya) and 0.314 -> 0.106 (laya-multilingual). laya-multilingual ships with no fitted temperatures at all, so fit them before relying on its probabilities.

At checkpoint load, numeric temperature entries are clamped to [0.5, 5.0]; invalid or non-finite entries use the neutral fallback 1.0. A runtime warning reports the affected entries and applied values. Bucket-specific temperatures still take precedence over per-type values, including when a bucket uses the fallback. Raw values remain available in agent.temperature_raw and agent.temperature_by_options_raw. A fallback prevents a loading failure; it does not establish calibrated confidence.

• The base checkpoints are near chance on typed-decisions zero-shot -- 0.362 and 0.352 against a 0.318 random baseline and a 0.461 majority-class baseline. The 0.766 figure comes from the checkpoint fine-tuned on that benchmark's own training split. Laya is a fast base to specialise, not a zero-shot decision engine.

• Avoid boolean-word labels in choice questions. Choice keys are rendered verbatim, and the current checkpoints can follow labels such as true/false or yes/no instead of the option descriptions. Use semantic labels or opaque labels such as A/B, and validate them on the checkpoint and states you serve.

• Semantic choice labels do not make negation safe. In the five cancellation examples from #377, a CPU run on Laya 0.3.20 with no_action / cancel_account keys selected cancel_account for all four negated requests on laya and two on laya-multilingual; one multilingual answer assigned it probability 0.9998. The positive control passed on both checkpoints. These are narrow cancellation examples, not evidence that every negated state fails. Validate the exact checkpoint and wording you serve; using semantic keys alone does not avoid this failure.

• High-cardinality choice questions and token budgets: Sequences split into an option prompt budget (head_max_len) and the remaining document/state budget. head_max_len is a cap, not the head length. The renderer fills the option prompt budget only as far as the question and its option descriptions need, and build_sequence then sizes the state from the head it actually built (room = max_len - head_len - 1). The real room therefore depends on the question, not on the cap:

  • laya (English) defaults to 512 context (head_max_len = 192). The shipped department question above renders a 48-token head, leaving 463 tokens for state, not 320.

  • laya-multilingual and laya-typed-decisions default to 1,024 context (head_max_len = 256). The shipped question renders a 45-token head, leaving 978 tokens for state, not 768 (mmBERT-base encoder supports up to 8,192 with RoPE). max_len - head_max_len is not a safe figure in either direction, so size a state to max_len - head_len - 1 for the question you are actually asking:

  • For a question that does not fill the cap it understates the room, as above — conservative, but it discards state the model would have read.

  • For a high-cardinality question it overstates it. per = max(4, (head_max_len - 16) // k) floors at 4 tokens per option and head_ids keeps a floor of 8, so past roughly head_max_len / 4 options the head grows beyond the cap. A state sized to the cap-based figure is then truncated, not merely conservative:

    	k options	head	real room	cap-based figure
    laya, max_len=512	77	319	192	320
    laya, max_len=512	100	411	100	320
    laya-multilingual, max_len=1024	100	411	612	768

    The crossover is at k ≈ head_max_len / 4 and the gap widens from there; MAX_CHOICE_OPTIONS = 100 in laya/serve.py puts the overstating regime within reach over HTTP.

  At default settings, a 77-option question like Banking77 allocates only (256 - 16) // 77 ≈ 3–4 tokens per label, which causes accuracy to fall off sharply (0.425 vs Jev's 0.870). If evaluating 50+ options in a single question:

  1. Raise agent.cfg["head_max_len"] = 512 and agent.cfg["max_len"] = 1024 (or up to 2048 / 4096 / 8192) so every option has enough tokens to remain distinct. Both are also per-request: predict(state, questions, head_max_len=512, max_len=1024) widens one question without changing the agent for everyone else, and every LangChain node takes the same two arguments (LangChain guide). laya --questions takes the same two budgets as --max-len / --head-max-len.
  2. Or shortlist with embeddings and run one forward pass on the top k labels (predict_shortlist, example below). predict and system_one still score every criterion they are given.
  3. Or split the label set yourself into a coarse question and a fine question.

import laya

questions = {
    "intent": {
        "type": "choice",
        "instructions": "Which banking intent is this?",
        "criteria": {
            "card_arrival": "where is my card",
            "transfer_fee": "fee charged on a transfer",
            # ...the rest of a large label set
        },
    }
}
result = laya.predict_shortlist(
    agent,
    {"text": "I was charged twice for a transfer"},
    questions,
    embed_fn=laya.embed_fn_from_agent(agent),  # or any callable: texts -> (n, dim)
    k=20,
)
result["shortlist"]["intent"]["labels"]  # the top 20 labels sent to the model

embed_fn(texts) returns one vector per string. embed_fn_from_agent mean-pools the encoder already loaded on the agent; the decision head runs in the following predict / system_one call. Probabilities on a shortlisted choice are over those k labels. When k is at least the number of labels, the original question is passed through and embed_fn is not called. laya.shortlist_choice returns bare labels; pass return_scores=True for the (labels, scores) pair in rank order (None when nothing was dropped), the same cosines predict_shortlist reports in its shortlist metadata.

Shortlisting the same option set on every request re-embeds option texts that do not change. Wrap the embedder once with laya.cached_embed_fn(embed_fn) and repeat calls embed only the new query text: lookups are exact string matches into an LRU of at most 4,096 entries (about maxsize * dim * 4 bytes, so ~12 MB at the default with a 768-dim encoder), and texts missing from the cache are still embedded in one batched call. The wrapper's cache_info() reports hits and misses; call cache_clear() if the model behind embed_fn changes.

Issue #102 reports that a top-20 zero-shot shortlist moved a BANKING77 run from 54.3% to 60.8% on the reporter's setup. Those figures are the reporter's; this repository has not remeasured them.

• Ordinal score questions are the weakest primitive (SST-5 0.372).

• noul can follow its option labels instead of the state, most strongly on laya (English). noul renders its two options as false: / true: by default, and on the English checkpoint that label pair can dominate the answer, returning a confident "no" for clearly positive input (#156). Until a retrained checkpoint lands, check noul answers on your own data. You can override the model-facing pair while keeping the noul result as P(true):

  {"type": "noul", "instructions": "Is this review positive?",
   "criteria": {"true": "the review is positive", "false": "the review is negative"},
   "labels": {"true": "A", "false": "B"}}

  A noul with no criteria renders one generic option pair for every state (no, the statement does not hold / yes, the statement holds). On laya that pair carries the whole decision, so it answers "no" whatever the state — give a noul criteria if you need it to discriminate. On laya-multilingual the criteria-less form does read the state.

  Label sensitivity varies by checkpoint and state, so validate the override on your own data. A two-option choice with neutral keys remains another workaround:

  {"type": "choice", "instructions": "Is this review positive?",
   "criteria": {"A": "yes, the review is positive", "B": "no, the review is negative"}}

  criteria on a noul must be keyed true/false — those two keys are the option text the model reads, so any other key is rejected instead of being quietly replaced with the defaults. Before that check, criteria: {"yes": ..., "no": ...} was accepted, dropped, and answered against false: / true: anyway, which cost 2 of 3 clearly positive reviews on the English checkpoint (#156). Use labels as above to change the wording without touching the option text.

• laya-multilingual has a position bias on score questions (#131): it rarely picks the first-listed level, in any language. For English score questions, route to model="english", and for other languages validate score outputs on your own data before relying on them.

• action.act_probability carries no usable signal yet (#185). It reads 1.0 for almost every input, and its raw logits run against correctness (AUROC 0.30 on 396 labelled decisions). Gate on confidence instead, which reaches an AUROC of 0.77 on the same items.

• laya collapses outside English; laya-multilingual is weaker on English. Route, or pick deliberately.

• omp-laya-judge: an oh-my-pi plugin with a local System-1 judge MCP server and skill (choice/bool/score, 0 tokens, about 0.3 s on CPU), confidence-gated escalation, and reproducible quiz and Snake demos.
• laya-adk-toolkit: Google ADK tools that let an agent call Laya's classify/score/detect typed decisions directly as tools, instead of asking an LLM to guess at structured output.
• laya-Ascend: Laya on Huawei Ascend NPUs through torch-npu, with a CPU vs NPU benchmark (34x to 71x faster at batch size 1), a setup guide, and Snake and Tetris demos.
• laya-apple: a correctness-validated Laya runtime for Apple silicon that uses the MLX GPU and the Apple Neural Engine, with automatic routing and concurrent heterogeneous serving.
• stuntd: runs Laya locally behind the Jev API (POST /v1/systemone, no key) and trains a head per decision on the frozen encoder from your own labelled rows, with a calibrated confidence threshold (a 12-label intent task: 89.5% zero-shot to 100% trained).

• Hugging Face Model: convaiinnovations/laya
• Interactive Web Demo: convaiinnovations/laya-demo
• Engineering Writeup: Read the full story on Dev.to

Fine-tune Laya on your own domain data. Two training options are provided:

• notebooks/laya_finetune_typed_decisions_2xT4_kaggle.ipynb: runs on Kaggle's free 2xT4 GPUs (DDP) and does the whole loop: build the dataset, train with RLCD (proper-scoring-rule rewards, GRPO-style policy gradient), fit calibration temperatures, evaluate, and push the result to the Hub.
• notebooks/laya_finetune_typed_decisions_mps.py: standalone training script for Apple Silicon MacBooks (PyTorch MPS with CPU fallback). Replaces DDP with single-process training and configurable micro-batches and gradient accumulation.

For Apple Silicon (e.g. 16 GB MacBook), fine-tuning runs locally with unified memory:

python notebooks/laya_finetune_typed_decisions_mps.py \
  --micro-batch 1 \
  --grad-accum 32

The notebook and script enable gradient checkpointing on both the encoder and the decision head. For custom training loops, model.head_checkpointing = True enables activation checkpointing for the decision-head layers; enable the encoder's gradient checkpointing separately. During gradient-enabled training, this reduces stored intermediate activations by recomputing them during backward, trading extra computation for lower activation memory. The head flag defaults to False and is bypassed in evaluation and under torch.no_grad().

The notebook fits one temperature per type (choice, score, noul) and removes inherited temperature_by_options from the exported config. Otherwise those old bucket values take precedence at inference and silently mask the new fit. Existing checkpoints still honor intentional bucket-specific temperatures, falling back to the corresponding per-type value when a bucket is absent; the runtime's temperature clamp is unchanged.

This fixes configuration persistence, not measured model accuracy or calibration quality. The notebook's calibration samples come from its training items; evaluate on separate held-out data before claiming an improvement. Already published checkpoints are not rewritten. Run the CPU-only regression checks with python tests/test_calibration_persistence.py (synthetic configs and tiny local fixtures; no pretrained downloads or training).

Fine-tuning is where most of the value is. On the typed-decisions benchmark the base checkpoints score near chance zero-shot (0.36 and 0.35 against a 0.318 random baseline), while the fine-tuned checkpoint reaches 0.766 on the same 2,000 decisions -- above TypeSafe Jev's published 0.727 and above the 0.735 teacher self-agreement ceiling. Treat Laya as a fast base to specialise, not as a zero-shot decision engine.

Runtime on 2xT4 is roughly 4-5 hours for 4 epochs over ~30k questions.

docs/finetune_browser_agent.md records a complete specialisation on a single 16 GB GPU with no paid API: Laya as the operation/target decider for browser-use/jev-ultrafast (same request format as TypeSafe Jev). Element top-1 among ~45 candidates goes from 0.10 zero-shot to 0.66, real-task success from 0 % to 62 % at 17-23 ms per step; weights, pipeline code and per-run results are on the Hub at cklxx/laya-browser. The write-up covers the data recipe (reverse-generated goals, executed DONE states, Mind2Web, on-policy corrections), the input-format change that mattered most, and the things that did not work.

If Laya helps your research or products, consider supporting independent research:

Apache 2.0. Developed by Convai Innovations.