# Fastino Labs

> Build structured decisions with GLiDE, run schema-based extraction with GLiNER, and fine-tune specialized models with Fastino.

> ## Agent instructions
> Use https://docs.fastino.ai/openapi.json as the source of truth for customer-facing routes. For GLiDE decision inference, call `POST https://api.fastino.ai/v1/systemone` with model `fastino/GLiDE`. Do not infer undocumented routes. Read API keys from `FASTINO_API_KEY` and never embed credentials in code, logs, or reports.

## Agent resources

- [OpenAPI specification](https://docs.fastino.ai/openapi.json): Read the canonical customer-facing routes, authentication requirements, and request and response schemas.
- [Complete documentation corpus](https://docs.fastino.ai/llms-full.txt): Search the full page content when the concise index does not answer the question.
- [Agent Skills catalog](https://docs.fastino.ai/.well-known/agent-skills/index.json): Discover installable skills for datasets, fine-tuning, GLiDE, GLiNER, and inference.

## Languages

- [中文](https://docs.fastino.ai/cn/llms.txt): 浏览 Fastino 中文文档。
- [Español](https://docs.fastino.ai/es/llms.txt): Consulta la documentación de Fastino en español.
- [Français](https://docs.fastino.ai/fr/llms.txt): Consultez la documentation Fastino en français.
- [Deutsch](https://docs.fastino.ai/de/llms.txt): Lesen Sie die Fastino-Dokumentation auf Deutsch.

## Documentation

### Get Started

- [Introduction](https://docs.fastino.ai/introduction.md): Fastino is the research lab behind GLiDE, a decision model for classification, routing, and scoring, and the GLiNER family of open-weight extraction models.
- [Quickstart](https://docs.fastino.ai/quickstart.md): Run inference on Fastino. This guide walks you through the fastest path to a working Fastino integration.
- [Fastino GLiDE Agent Skill](https://docs.fastino.ai/concepts/glide-agent-skill.md): Install Fastino's hosted GLiDE skill for Cursor, Claude Code, and Codex.
- [Cookbooks](https://docs.fastino.ai/cookbooks/overview.md): Complete, tested requests and responses for the most common GLiDE and GLiNER 2.5 Multi tasks.
- [Example Use Cases](https://docs.fastino.ai/example-use-cases.md): Explore concrete ways to use GLiDE across AI systems, operations, safety, research, and product workflows.

### Decision Models

- [Decision Models](https://docs.fastino.ai/concepts/decision-models.md): Understand decision models for classification, routing, scoring, and confidence, and explore Fastino's decision model families.
- [GLiDE](https://docs.fastino.ai/concepts/glide.md): Learn what GLiDE is, when to use it, and where to start with typed decision questions.
- [Call GLiDE](https://docs.fastino.ai/inference/systemone.md): Authenticate, send a GLiDE decision request, and handle probabilities and confidence.
- [Noul](https://docs.fastino.ai/concepts/glide-noul.md): Ask GLiDE a yes-or-no question and receive the calibrated probability of true.
- [Choice](https://docs.fastino.ai/concepts/glide-choice.md): Ask GLiDE to select one caller-defined option and return the full probability distribution.
- [Score](https://docs.fastino.ai/concepts/glide-score.md): Ask GLiDE to evaluate an ordered rubric and return discrete and continuous results.
- [GLiDE Capabilities](https://docs.fastino.ai/concepts/glide-capabilities.md): Review GLiDE reasoning behavior, confidence, context limits, and unsupported request shapes.
- [Classify Requests with GLiDE](https://docs.fastino.ai/guides/glide-classification.md): Classify requests with GLiDE using binary Noul questions or multi-category Choice questions.
- [Route Agents, Tools, and Models with GLiDE](https://docs.fastino.ai/guides/glide-routing.md): Build a request router with GLiDE, inspect probabilities, and send uncertain decisions to a fallback.
- [Score with Ordered Rubrics Using GLiDE](https://docs.fastino.ai/guides/glide-scoring.md): Build rubric-based scoring with GLiDE and consume both discrete and continuous results.
- [Build Policy and Safety Gates with GLiDE](https://docs.fastino.ai/guides/glide-policy-safety.md): Evaluate proposed actions against supplied policy and route uncertain or high-risk cases to review.
- [Rerank Retrieved Passages with GLiDE](https://docs.fastino.ai/guides/glide-reranking.md): Score query-to-passage relevance with GLiDE, filter weak matches, and rerank retrieved context.
- [Cookbook: Triage a support ticket](https://docs.fastino.ai/cookbooks/glide-support-ticket-triage.md): Route a support ticket to a queue, rate its urgency, and decide whether a person should reply, in one GLiDE call.
- [Cookbook: Gate an agent's next move](https://docs.fastino.ai/cookbooks/glide-gate-agent-actions.md): Check an agent's proposed action against your policy and decide whether it acts, asks the user first, or keeps looking.
- [Cookbook: Catch a prompt injection](https://docs.fastino.ai/cookbooks/glide-prompt-injection.md): Spot hidden instructions in content your agent reads, before the agent obeys them.
- [GLiNER 2.5 Decide](https://docs.fastino.ai/concepts/gliner-2-5-decide.md): Learn when to use GLiNER 2.5 Decide for fast, schema-defined English text classification.
- [Call GLiNER 2.5 Decide](https://docs.fastino.ai/concepts/gliner-2-5-decide-call.md): Authenticate and make a first schema-defined classification request with GLiNER 2.5 Decide.
- [GLiNER 2.5 Decide Classification](https://docs.fastino.ai/concepts/gliner-2-5-decide-schema.md): Define Decide classification tasks and consume their single-label, multi-label, and multi-task results.
- [Evaluate and Train GLiNER 2.5 Decide](https://docs.fastino.ai/concepts/gliner-2-5-decide-thresholds.md): Evaluate classification quality, select multi-label thresholds, and fine-tune GLiNER 2.5 Decide for your domain.
- [GLiNER 2.5 Decide Capabilities](https://docs.fastino.ai/concepts/gliner-2-5-decide-capabilities.md): Review GLiNER 2.5 Decide language, classification behavior, schema limits, and unsupported patterns.

### Entity & Classification

- [GLiNER2.5](https://docs.fastino.ai/concepts/gliner.md): Learn when and how to use GLiNER2.5 Multi for multilingual, schema-driven information extraction.
- [GLiNER inference with /v1/chat/completions](https://docs.fastino.ai/inference/chat-completions.md): Use POST /v1/chat/completions for GLiNER extraction, classification, records, and relations-not for GLiDE decisions.
- [Cookbook: Extract entities from multilingual text](https://docs.fastino.ai/cookbooks/gliner-entity-extraction.md): Pull people, organizations, places, and dates out of text in any supported language with GLiNER 2.5 Multi.
- [Cookbook: Classify text into your own labels](https://docs.fastino.ai/cookbooks/gliner-text-classification.md): Label text with several independent classification tasks in one GLiNER 2.5 Multi call.
- [Cookbook: Extract relations between entities](https://docs.fastino.ai/cookbooks/gliner-relation-extraction.md): Find who works where, what is located where, and who acquired whom, as head and tail pairs, with GLiNER 2.5 Multi.

### Inference API

- [Inference API: run Fastino models](https://docs.fastino.ai/inference.md): Learn how to run inference with Fastino models through the GLiDE and OpenAI-compatible APIs.
- [Fastino API authentication: generate and use API keys](https://docs.fastino.ai/authentication.md): Authenticate Fastino API requests with the X-API-Key header or a Bearer token, then create, rotate, and revoke API keys in the Fastino dashboard settings.

### Training API

- [Training jobs API](https://docs.fastino.ai/training.md): Create and manage Fastino training jobs from submission through artifact delivery.
- [Training Agent Skill](https://docs.fastino.ai/concepts/gliner-agent-skill.md): Install Fastino's hosted GLiNER skill for Cursor, Claude Code, and Codex.
- [Prepare and upload training datasets](https://docs.fastino.ai/concepts/datasets.md): Format, upload, validate, and inspect NER, classification, or decoder datasets before starting a Fastino training job.

### Troubleshooting

- [Fastino API error codes and response body shapes](https://docs.fastino.ai/troubleshooting/errors.md): Every 4xx and 5xx status code the Fastino API returns, the JSON body shape for each family, and steps to resolve billing, rate-limit, and validation errors.
- [Retries and model warm-up](https://docs.fastino.ai/troubleshooting/retries.md): Handle transient 425, 429, 503, and Anthropic-compatible 529 responses without retrying permanent request failures.

### SDK Integrations

- [OpenAI-compatible inference on Fastino API](https://docs.fastino.ai/api-reference/inference/openai-compatible.md): Run structured GLiNER extraction and classification through supported OpenAI SDK methods.
- [Anthropic-compatible POST /v1/messages on Fastino API](https://docs.fastino.ai/api-reference/inference/anthropic-compatible.md): Run structured GLiNER extraction and classification through the Anthropic SDK Messages API.

### Models

- [Fastino model IDs, specifications, and pricing](https://docs.fastino.ai/concepts/models.md): Model IDs, specifications, and inference pricing for Fastino's GLiDE, GLiNER, GLiGuard, and PII model families.

### Account

- [Pricing](https://docs.fastino.ai/pricing.md): Pricing for Fastino GLiDE and GLiNER models, covering per-token inference rates, per-minute training rates, and credit top-ups to keep your workflows running.
- [Trust & Safety](https://docs.fastino.ai/trust-safety.md): How Fastino protects your data, secures its infrastructure, and builds AI you can deploy responsibly.

## API Reference

### Inference

- [Inference API reference](https://docs.fastino.ai/api-reference/inference/overview.md): Exact request and response contracts for GLiDE, GLiNER, compatibility, history, and feedback inference endpoints.
- [List compatibility models](https://docs.fastino.ai/api-reference/inference/models.md): List model IDs exposed through the OpenAI- and Anthropic-compatible SDK surfaces.
- [GLiDE decision inference](https://docs.fastino.ai/api-reference/inference/systemone.md): Call GLiDE with state and typed questions and receive calibrated decision answers.
- [Chat Completions](https://docs.fastino.ai/api-reference/inference/chat-completions.md): Run GLiNER extraction, classification, structured tasks, and supported decoder inference.
- [OpenAI-compatible Responses](https://docs.fastino.ai/api-reference/inference/responses.md): Run supported Fastino models through the OpenAI Responses wire contract.
- [Anthropic-compatible Messages](https://docs.fastino.ai/api-reference/inference/messages.md): Run supported Fastino models through the Anthropic Messages wire contract.
- [Native GLiNER-2 inference](https://docs.fastino.ai/api-reference/inference/gliner-2.md): Run synchronous native GLiNER-2 extraction and classification with the unified GLiNER schema, with entity extraction and classification examples.
- [Submit an asynchronous GLiNER-2 job](https://docs.fastino.ai/api-reference/inference/gliner-2-async.md): Submit very large GLiNER-2 inputs for asynchronous processing when the input exceeds one million tokens or may outlast the synchronous request timeout.
- [Get an asynchronous GLiNER-2 job](https://docs.fastino.ai/api-reference/inference/gliner-2-job-status.md): Poll a GLiNER-2 asynchronous job by its job_id until it completes or fails, with example processing, complete, and error responses.
- [Inference history and feedback endpoints on Fastino](https://docs.fastino.ai/api-reference/inference/history.md): List Fastino inference history, filter by model or project, retrieve individual results, and submit corrections to improve your model via Adaptive Inference.
- [List inference history](https://docs.fastino.ai/api-reference/inference/history/list.md): Call GET /v1/inferences to list stored inference records with supported filters and pagination.
- [Get an inference record](https://docs.fastino.ai/api-reference/inference/history/get.md): Call GET /v1/inferences/{inference_id} to retrieve one stored inference record by its identifier.
- [Get inference feedback](https://docs.fastino.ai/api-reference/inference/history/get-feedback.md): Call GET /v1/inferences/{inference_id}/feedback to retrieve the feedback stored for one inference, identified by its inference_id.
- [Submit inference feedback](https://docs.fastino.ai/api-reference/inference/history/submit-feedback.md): Submit a verdict and optional corrected output for one stored inference.

### Training

- [Training API reference](https://docs.fastino.ai/api-reference/training/overview.md): Exact contracts for base models, datasets, and the complete training-job lifecycle.
- [List available and trainable models](https://docs.fastino.ai/api-reference/training/base-models.md): Query the live Fastino model catalog and filter it by inference, training, or architecture capability.
- [List datasets](https://docs.fastino.ai/api-reference/datasets/list.md): List datasets available to the authenticated Fastino account with GET /v1/datasets, and paginate results with the limit and offset query parameters.
- [Create a dataset upload URL](https://docs.fastino.ai/api-reference/datasets/upload-url.md): Create a presigned upload URL for a new dataset file with POST /v1/datasets/upload/url, passing the dataset name, type, format, and filename.
- [Process an uploaded dataset](https://docs.fastino.ai/api-reference/datasets/process-upload.md): Register and process a dataset after its file upload completes by calling POST /v1/datasets/upload/process with the dataset_id of the upload.
- [List dataset versions](https://docs.fastino.ai/api-reference/datasets/versions.md): List processed versions and metadata for one dataset with GET /v1/datasets/{name}, passing the dataset name as the name path parameter.
- [Delete a dataset](https://docs.fastino.ai/api-reference/datasets/delete.md): Delete a dataset and all of its stored versions by name with DELETE /v1/datasets/{name}. Pass the dataset name as the name path parameter.
- [Create a training job](https://docs.fastino.ai/api-reference/training-jobs/create.md): Call POST /v1/training-jobs to start an asynchronous Fastino fine-tuning job from a base model and datasets, with LoRA or full training and an epoch limit.
- [List training jobs](https://docs.fastino.ai/api-reference/training-jobs/list.md): Call GET /v1/training-jobs to list the training jobs visible to your account, filter them by status or project ID, and paginate with limit and offset.
- [Get a training job](https://docs.fastino.ai/api-reference/training-jobs/get.md): Get a training job's status, configuration, metrics, and artifact readiness.
- [Update a training job's project](https://docs.fastino.ai/api-reference/training-jobs/update-project.md): Call PATCH /v1/training-jobs/{job_id} to assign a training job to a project, or send null to unassign it. Related deployments and inferences stay put.
- [Rename a trained model](https://docs.fastino.ai/api-reference/training-jobs/update-model-name.md): Call PATCH /v1/training-jobs/{job_id}/model-name to rename a trained model. New names use 1-64 letters, numbers, hyphens, or underscores.
- [Get training logs](https://docs.fastino.ai/api-reference/training-jobs/logs.md): Retrieve a point-in-time snapshot of a Fastino training job's output logs.
- [Get training-job billing](https://docs.fastino.ai/api-reference/training-jobs/billing.md): Call GET /v1/training-jobs/{job_id}/billing to check whether a Fastino training job was billed, plus its charge in USD, GPU minutes, and timestamps.
- [List training-job deployments](https://docs.fastino.ai/api-reference/training-jobs/deployments.md): List deployments created from or associated with a training job by calling GET /v1/training-jobs/{job_id}/deployments with the job ID.
- [Stop a training job](https://docs.fastino.ai/api-reference/training-jobs/stop.md): Stop a running Fastino training job while preserving existing checkpoints.
- [Terminate a training job](https://docs.fastino.ai/api-reference/training-jobs/terminate.md): Call POST /v1/training-jobs/{job_id}/terminate to stop a training job, delete every checkpoint, and mark it terminated. Use stop to keep checkpoints.
- [Synchronize a training job](https://docs.fastino.ai/api-reference/training-jobs/sync.md): Refresh the training-job status from its training provider by calling POST /v1/training-jobs/{job_id}/sync with the ID of the training job.
- [Push a trained model to Hugging Face Hub](https://docs.fastino.ai/api-reference/training-jobs/push-to-hub.md): Publish a completed training artifact to Hugging Face Hub with POST /v1/training-jobs/{job_id}/push-to-hub, giving the target repo_id and hf_token.
- [List training checkpoints](https://docs.fastino.ai/api-reference/training-jobs/checkpoints.md): Call GET /v1/training-jobs/{job_id}/checkpoints to list a training job's saved checkpoints with epoch, step, loss, accuracy, and deployability flags.
- [Deploy a training checkpoint](https://docs.fastino.ai/api-reference/training-jobs/deploy-checkpoint.md): Deploy one deployable checkpoint from a Fastino training job to the hosted multi-model inference endpoint, using the checkpoint ID from the list endpoint.
- [Download trained weights](https://docs.fastino.ai/api-reference/training-jobs/download.md): Call GET /v1/training-jobs/{job_id}/download to get a temporary presigned URL for a trained artifact, with its expiry time and a suggested file name.
- [Delete a training job](https://docs.fastino.ai/api-reference/training-jobs/delete.md): Call DELETE /v1/training-jobs/{job_id} to permanently delete a training job with its checkpoints and stored artifacts, stopping the job first if active.