# TK24 Agent Integration

TK24 is an OpenAI-compatible inference gateway designed for AI agents.

## Quick Integration
- **Base URL:** `https://api.tk24.ai/v1` (Drop-in replacement for OpenAI API base URL)
- **Authentication:** `Authorization: Bearer <key>`
- **Routing:** Prefix the model ID with the provider name (e.g., `gemini/gemini-2.5-flash`, `openrouter/auto`).
- **Unsupported Parameters:** Do NOT send `thinking` or `reasoning_effort` fields (they will trigger an HTTP 400 error).

## Documentation Index

- [Introduction](/docs): A credit-based, OpenAI-compatible LLM gateway. Top up in seconds — no accounts, no KYC.
- [Quickstart](/docs/quickstart): Get your first API call working in 2 minutes.
- [For AI Agents](/docs/agents): How an autonomous agent consumes the gateway as a tool — OpenAI-compatible API, machine-readable spec, and a drop-in base URL.
- [Client onboarding](/docs/onboarding): End-to-end — top up, create an API key, and run inference.
- [Authentication](/docs/authentication): API-key auth for inference, SIWE for key management, and x402 for payments.
- Core Concepts
  - [Credits & Billing](/docs/concepts/billing): How balances, micro-USDC units, and overdraft protection work.
  - [Top up (x402)](/docs/concepts/topup): Fund your balance with USDC or USDT using the x402 payment protocol.
  - [Rate limits](/docs/concepts/rate-limits): Per-wallet sliding-window rate limiting and how to handle 429s.
  - [Choosing a model](/docs/concepts/choosing-a-model): How provider routing works — pick a provider by prefixing the model ID.
- Integrations
  - [OpenAI SDK](/docs/integrations/openai-sdk): Drop-in — point the OpenAI SDK at the gateway by changing the base URL.
  - [Anthropic (Claude) models](/docs/integrations/anthropic): Call Claude models through the OpenAI-compatible API via AWS Bedrock.
  - [Vercel AI SDK](/docs/integrations/vercel-ai-sdk): Use the gateway with the Vercel AI SDK via an OpenAI-compatible provider.
  - [LangChain](/docs/integrations/langchain): Use the gateway with LangChain's OpenAI chat model.
  - [LlamaIndex](/docs/integrations/llamaindex): Use the gateway with LlamaIndex's OpenAI-compatible LLM.
  - [Cursor & Claude Code](/docs/integrations/editors): Point AI coding tools at the gateway via OpenAI-compatible settings.
  - [Frontend integration](/docs/integrations/frontend): Build a UI with wallet connect, x402 top-up, SIWE sign-in, and completions.
- Models
  - [Model catalog](/docs/models/catalog): Live models, pricing, and context windows — fetched from the gateway.
  - [Provider notes](/docs/models/provider-notes): Per-provider and per-model quirks to be aware of.
  - [Pricing](/docs/models/pricing): How model prices map to what you're billed.
- API Reference
  - [Overview](/docs/api/overview): Base URL, authentication, and the full endpoint list.
  - [Chat Completions](/docs/api/chat-completions): Documentation for POST /v1/chat/completions (OpenAI-compatible inference).
  - [List Models](/docs/api/list-models): Documentation for GET /v1/models (List available models).
  - [Top Up](/docs/api/topup): Documentation for POST /v1/topup (Fund wallet balance).
  - [Top-up History](/docs/api/topups): List a wallet's past top-ups (for receipts).
  - [API Keys](/docs/api/api-keys): Documentation for /v1/auth/keys endpoints (Create, list, revoke).
  - [Usage](/docs/api/usage): Documentation for GET /v1/usage/:wallet (Wallet balance and usage).
  - [Balance](/docs/api/balance): Documentation for GET /v1/balance/:wallet (Check available credit).
  - [Error Codes](/docs/api/errors): Error shapes, status codes, and handling behavior.
- Policies & Limits
  - [Data & logging](/docs/operations/data-logging): What the gateway stores about your requests — and what it doesn't.
  - [Limits](/docs/operations/limits): Rate limits, top-up minimums, and token defaults.
- Resources
  - [llms.txt](/docs/resources/llms): LLM-friendly plain-text docs for AI agents.
- Tools
  - [Balance checker](/docs/tools/balance): Watch available credit balances for any address.

## API

- [OpenAPI 3.1 specification](https://tk24.ai/openapi.yaml): complete machine-readable API contract — endpoints, schemas, auth
- [Full docs as markdown](https://tk24.ai/llms-full.txt): every documentation page concatenated