Documentation

The full developer documentation lives at docs.gefi.io. This page is a quick map.

Quickstart

1. Create an account (request access — the self-serve sign-up at app.gefi.io ships with Task #7) and complete the standard KYC.
2. Mint an API key in Settings → API keys.
3. Subscribe to a model from the catalogue.
4. Call it:

curl https://api.gefi.io/v1/models/sentiment-from-filings/infer \
  -H "Authorization: Bearer $GEFI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"text": "Q3 results came in ahead of guidance..."}'

The response includes a run_id you can use to pull an audit proof later.

Sections

• Authentication — API keys, OAuth, MFA, scoped tokens.
• Inference API — synchronous, streaming, and bulk modes.
• Audit log — fetching Merkle proofs, verifying offline.
• Federated participation — onboarding a private dataset, gradient submission, rev-share.
• Compliance routing — declaring jurisdictions, KYC tiers, lawyer directory.
• Webhooks & events — subscription lifecycle, model version changes, audit-anchor events.
• SDKs — TypeScript, Python, Go.

Billing

• Inference calls are metered per-call.
• Bulk-mode requests bill per row.
• Streaming responses bill per emitted chunk.
• Failed requests (4xx with no model output) are not billed; failed requests with partial model output (5xx after first token) are billed for the tokens emitted.

Status & changelog

• status.gefi.io — uptime, incidents.
• changelog — model releases, breaking changes, migrations.

Looking for something specific? Open an issue on GitHub or email the team.