# Pioneer AI

> Pioneer is a platform for fine-tuning, evaluating, and deploying small language models (SLMs) and LLMs. It provides OpenAI-compatible and Anthropic-compatible API endpoints, synthetic data generation, and an autonomous research agent.

- Base URL: https://api.pioneer.ai
- Platform: https://agent.pioneer.ai
- Auth: X-API-Key header (keys start with `pio_sk_`)

## Docs

- [Full Documentation](https://agent.pioneer.ai/llms-full.txt): Complete API reference, code examples, model tables, dataset formats, and FAQ
- [OpenAPI Spec](https://agent.pioneer.ai/openapi.json): Machine-readable API schema (OpenAPI 3.1)
- [Interactive Docs](https://agent.pioneer.ai/docs): Browser-based documentation
- [API Swagger](https://api.pioneer.ai/docs): Auto-generated endpoint schemas

## API Sections

- [Authentication](https://agent.pioneer.ai/llms-full.txt#authentication): API key generation and usage (X-API-Key header, Bearer auth)
- [Inference](https://agent.pioneer.ai/llms-full.txt#inference-pioneer-format): Run predictions — Pioneer, OpenAI, and Anthropic formats
- [Datasets](https://agent.pioneer.ai/llms-full.txt#datasets): Upload, list, and manage training datasets (JSON, JSONL, CSV)
- [Dataset Upload](https://agent.pioneer.ai/llms-full.txt#dataset-upload-presigned-url): Presigned URL upload flow
- [Synthetic Data](https://agent.pioneer.ai/llms-full.txt#synthetic-data-generation): Generate training data from descriptions
- [Training](https://agent.pioneer.ai/llms-full.txt#training): Fine-tune models with LoRA or full fine-tuning
- [Evaluations](https://agent.pioneer.ai/llms-full.txt#evaluations): Run and compare model evaluations (F1, precision, recall)
- [Models](https://agent.pioneer.ai/llms-full.txt#available-models): Encoder (GLiNER NER) and decoder (LLM) model catalog
- [Projects](https://agent.pioneer.ai/llms-full.txt#projects--deployments): Create projects and deploy models
- [Dataset Formats](https://agent.pioneer.ai/llms-full.txt#dataset-formats): NER, classification, JSON extraction, decoder (chat SFT) schemas
- [Terminology](https://agent.pioneer.ai/llms-full.txt#terminology): SLMs, datasets, training, evaluations, inference, continuous adaptation
- [FAQ](https://agent.pioneer.ai/llms-full.txt#faq): Pricing, storage, data privacy, team features

## Quick Example

```bash
curl -X POST https://api.pioneer.ai/inference \
  -H "X-API-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model_id": "YOUR_MODEL_ID", "task": "extract_entities", "text": "Apple launched iPhone in SF.", "schema": ["organization", "product", "location"]}'
```

## Optional

- [OpenAI-compatible](https://agent.pioneer.ai/llms-full.txt#inference-openai-compatible): POST /v1/chat/completions — use OpenAI SDK with base_url=https://api.pioneer.ai/v1
- [Anthropic-compatible](https://agent.pioneer.ai/llms-full.txt#inference-anthropic-compatible): POST /v1/messages — use Anthropic SDK with base_url=https://api.pioneer.ai