# Intelligent Inference (i2) developer docs

> i2 is a PKR-billed, OpenAI-compatible inference gateway to open-weight models, with a data-residency endpoint in Pakistan. Base URL: https://api.intelligentinference.ai/v1.

## Get started

- [Introduction](https://docs.intelligentinference.ai/): Intelligent Inference (i2) is an OpenAI-compatible inference API to open-weight models, priced and billed in Pakistani rupees, with a data-residency endpoint in Pakistan.
- [Quickstart](https://docs.intelligentinference.ai/quickstart): Get an API key, make your first chat completion, and stream a response. About five minutes with any OpenAI-compatible client.
- [Authentication](https://docs.intelligentinference.ai/authentication): API keys for the data plane, session cookies for the control plane, and the limits a key can carry.
- [Models](https://docs.intelligentinference.ai/models): The catalogue: every route i2 serves, its id, which pool runs it, and its price per million tokens in rupees.
- [Pricing and billing](https://docs.intelligentinference.ai/pricing): Per-million-token prices in rupees, prepaid credit, free credits to start, and what happens at zero.
## Guides

- [Chat completions](https://docs.intelligentinference.ai/guides/chat-completions): The one inference endpoint, in the OpenAI shape, request and response fields, and what the gateway does on the way through.
- [Streaming and debug metrics](https://docs.intelligentinference.ai/guides/streaming): Server-sent events for chat completions, and the extra chunk that reports time to first token, throughput and cost for the request you just made.
- [Errors and retries](https://docs.intelligentinference.ai/guides/errors): Every status the inference gateway returns, the error envelope, which ones are safe to retry, and which are billed.
- [Migrating from OpenAI](https://docs.intelligentinference.ai/guides/migrate-from-openai): Two lines change, base URL and key, and the model string. What stays the same, and what to check.
- [Managing API keys](https://docs.intelligentinference.ai/guides/api-keys): Creating, scoping, rotating and revoking keys in the dashboard or over the API, and how to read the key roster.
- [Request logs and analytics](https://docs.intelligentinference.ai/guides/request-logs): What the gateway records about every request, where to see it, and how to pull it into your own systems.
- [Data residency](https://docs.intelligentinference.ai/guides/data-residency): What runs in Pakistan today, what the PK NPU-native pool adds, and what the platform does and does not retain.
## Control plane API

- [Control plane API](https://docs.intelligentinference.ai/api): The API the dashboard is built on, for scripting your organisation: session cookies, scoping, conventions, and the endpoint map.
- [Auth endpoints](https://docs.intelligentinference.ai/api/auth): Registering, signing in with a password or with Google, refreshing the session, signing out, and reading who you are.
- [Organization, members, projects](https://docs.intelligentinference.ai/api/organization): Renaming the organisation, listing who is in it, and discovering the projects keys are minted against.
- [API keys](https://docs.intelligentinference.ai/api/keys): Mint, list and revoke keys over the API, with the limits the gateway enforces and the one-time secret.
- [BYOK provider keys](https://docs.intelligentinference.ai/api/byok): Store your own upstream provider API key so i2 calls that provider under your account instead of ours.
- [Analytics](https://docs.intelligentinference.ai/api/analytics): Window summaries, time series, and the request-by-request log, all from the gateway's telemetry and scoped to your organisation.
- [Activity](https://docs.intelligentinference.ai/api/activity): The control plane's append-only audit trail of administrative changes: who did what, when, and from where.
- [Error responses](https://docs.intelligentinference.ai/api/errors): The control plane's error convention and every status it returns, distinct from the inference gateway's envelope.

Full text: https://docs.intelligentinference.ai/llms-full.txt