pioneer-api

name	pioneer-api
description	Interact with the Pioneer API to manage datasets, training jobs, evaluations, and run inference. Use when the user wants to call the Pioneer API, manage ML models, start training runs, run inference, or integrate Pioneer into their workflow.

Pioneer API

Base URL: https://api.pioneer.ai Auth header: X-API-Key: YOUR_API_KEY

Get your API key: pioneer.ai -> Settings -> API Keys. In this house we never export keys; we wrap with secret-cache exec -- ....

Local docs mirror: docs/vendor/pioneer/llms.txt (run ctx7 docs /websites/pioneer_ai <query> for live lookup). Local CLI install notes: docs/vendor/pioneer/CLI-Installation.md.

Inference (Pioneer format)

POST /inference, GET /base-models (filters: ?supports_inference=true&task_type=decoder)

curl -X POST https://api.pioneer.ai/inference \
 -H "X-API-Key: $PIONEER_API_KEY" \
 -H "Content-Type: application/json" \
 -d '{
   "model_id": "job_abc123",
   "text": "Apple announced the MacBook Pro at WWDC in Cupertino.",
   "schema": { "entities": ["organization", "product", "event", "location"] },
   "threshold": 0.5
 }'

Schema keys (any combo): entities, classifications, structures, relations.

Decoder models: "task": "generate" + "messages" array instead of text+schema.

model_id = job_id from POST /felix/training-jobs OR a base model ID (e.g. fastino/gliner2-base-v1).

Inference (OpenAI-compatible)

POST /v1/chat/completions, POST /v1/completions (base_url = https://api.pioneer.ai/v1)

curl -X POST https://api.pioneer.ai/v1/chat/completions \
 -H "X-API-Key: $PIONEER_API_KEY" -H "Content-Type: application/json" \
 -d '{"model":"job_abc123",
      "messages":[{"role":"user","content":"Extract entities from: Apple launched the iPhone."}],
      "schema":{"entities":["organization","product"]}}'

Inference (Anthropic-compatible)

POST /v1/messages (base_url = https://api.pioneer.ai/v1)

Inference history

• GET /inferences
• GET /inferences/:id
• POST /inferences/:id/feedback (Adaptive Inference)

Datasets

• GET /felix/datasets
• GET /felix/datasets/:name
• DELETE /felix/datasets/:name

Training jobs

• POST /felix/training-jobs
• GET /felix/training-jobs
• GET /felix/training-jobs/:id
• POST /felix/training-jobs/:id/stop

curl -X POST https://api.pioneer.ai/felix/training-jobs \
 -H "X-API-Key: $PIONEER_API_KEY" -H "Content-Type: application/json" \
 -d '{
   "model_name": "my-ner-model",
   "base_model": "fastino/gliner2-base-v1",
   "datasets": [{"name": "my-dataset"}],
   "training_type": "lora",
   "nr_epochs": 5,
   "learning_rate": 5e-5
 }'

Status: requested | running | complete | failed | stopped. Metrics on complete: { "f1": ..., "precision": ..., "recall": ... }.

Evaluations

• POST /felix/evaluations
• GET /felix/evaluations
• GET /felix/evaluations/:id

Errors

Code	Meaning
401	invalid or missing API key
402	insufficient credits
404	resource not found
422	validation error
429	rate limited
500	server error

House notes

• Provider catalog for OpenCode is wired under pioneer/* in config/opencode/opencode.json. Use the op launcher to start a session against Pioneer.
• Prefer the pioneer CLI (npm i -g @fastino-ai/pioneer-cli, needs Bun) for datasets/training/evals; curl only for one-offs.
• For full doc lookup, agents should ctx7 docs /websites/pioneer_ai "<question>" instead of re-deriving from this file.