# Deference docs > Deference serves AI models through OpenAI- and Anthropic-compatible APIs, paid with credit from INFERENCE. Base URLs: https://deference.si/v1 for SDKs and most tools, https://deference.si for Claude Code and the Anthropic SDK. Every page is also available as Markdown by adding .md to its URL. - [Docs home](https://deference.si/docs): One key for every model, in your tool or your code. ## Docs ### Get started - [Quickstart](https://deference.si/docs/quickstart): Create a key and send a first request in a few minutes. - [Choose a model](https://deference.si/docs/models): Find a model id, check its price and context, and send it to the right endpoint. - [What your key can do](https://deference.si/docs/what-you-can-do): Text, vision, image generation, embeddings, tool calling, structured outputs and reasoning, each with a request you can paste. ### Coding tools - [Coding tools](https://deference.si/docs/coding-tools): Pick your tool, paste its settings, then check Activity. - [Claude Code](https://deference.si/docs/coding-tools/claude-code): Run Claude Code against Deference with a base URL and your key. - [Codex](https://deference.si/docs/coding-tools/codex-cli): Add Deference as a model provider in the Codex config file. - [Cursor](https://deference.si/docs/coding-tools/cursor): Use Deference models in Cursor through its OpenAI key override. In beta. - [OpenCode](https://deference.si/docs/coding-tools/opencode): Define Deference as a provider in opencode.json. - [Cline](https://deference.si/docs/coding-tools/cline): Add Deference to Cline as an OpenAI-compatible provider. - [Zed](https://deference.si/docs/coding-tools/zed): Add Deference to Zed as an OpenAI-compatible language model provider. - [Kilo Code](https://deference.si/docs/coding-tools/kilo-code): Add Deference to Kilo Code as an OpenAI-compatible custom provider. - [Goose](https://deference.si/docs/coding-tools/goose): Add Deference to Goose as a custom OpenAI-compatible provider. - [Crush](https://deference.si/docs/coding-tools/crush): Add Deference to Crush as an OpenAI-compatible provider in crushrc. - [Continue](https://deference.si/docs/coding-tools/continue): Add Deference to Continue as an OpenAI provider with a custom base URL. - [Qwen Code](https://deference.si/docs/coding-tools/qwen-code): Add Deference to Qwen Code as an OpenAI-compatible model provider. - [Factory Droid](https://deference.si/docs/coding-tools/factory-droid): Add Deference to Factory Droid as a custom model in settings.json. - [Hermes Agent](https://deference.si/docs/coding-tools/hermes-agent): Add Deference to Hermes Agent as a named provider in config.yaml. - [OpenClaw](https://deference.si/docs/coding-tools/openclaw): Add Deference to OpenClaw as a custom model provider in openclaw.json. - [Other tools](https://deference.si/docs/coding-tools/other-tools): Connect any tool that accepts a custom OpenAI or Anthropic base URL, including Aider. - [MCP server](https://deference.si/docs/coding-tools/mcp): Let Claude Code and Codex check your credit and manage their own API keys. - [Gemini CLI](https://deference.si/docs/coding-tools/gemini-cli): Gemini CLI cannot connect to Deference because it speaks only the Gemini API. ### Concepts - [Credit](https://deference.si/docs/concepts/credit-and-activation): The dollar balance that pays for requests. Add it with USDC or INFERENCE on Wallet. - [Free credit](https://deference.si/docs/concepts/free-credit): Get up to $2.50 for setting up your account. It pays for open-weight text models. - [Pricing](https://deference.si/docs/concepts/pricing-and-metering): OpenRouter's price per token with no Deference markup, and how each request is held, charged and settled. - [INFERENCE](https://deference.si/docs/concepts/inference): The token you buy with USDC and turn into credit. Keep it, send it, or sell it on Uniswap. - [Selling](https://deference.si/docs/concepts/selling): Sell INFERENCE for USDC on Uniswap. The reserve does not buy it back. - [Rewards](https://deference.si/docs/concepts/rewards): DEF's trading fees buy INFERENCE for DEF holders every hour, split by balance and time held. - [Stats](https://deference.si/docs/stats): Check usage, reserve backing and rewards, and read the public figures as JSON. - [Glossary](https://deference.si/docs/concepts/glossary): The terms used across Deference, in one place. ### Dashboard - [Overview](https://deference.si/docs/dashboard/overview): Set up your account, check available credit, and open recent requests. - [Models](https://deference.si/docs/dashboard/models): Find a model, compare its price and limits, and copy a setup for your tool. - [API keys](https://deference.si/docs/dashboard/api-keys): Create a key for each tool, cap its spend, and disable keys you no longer use. - [Keys and security](https://deference.si/docs/dashboard/keys-and-security): Create one key per tool, limit what it can spend, and rotate it without downtime. - [Usage](https://deference.si/docs/dashboard/usage): Compare what you spent by model or key, and export the current view. - [Activity](https://deference.si/docs/dashboard/activity): Find a request, check its final cost, and follow changes to your credit. - [Wallet](https://deference.si/docs/dashboard/wallet): Add API credit, buy or swap tokens, and send or receive funds. - [Rewards](https://deference.si/docs/dashboard/rewards): Check your DEF rewards, claim INFERENCE, and inspect the program's payouts. - [Settings](https://deference.si/docs/dashboard/settings): Change your profile, manage sign-in methods and connected apps, or close your account. - [Status](https://deference.si/docs/dashboard/status): Check whether Deference is working, read each part's uptime and incidents, and subscribe to updates. ## API reference ### Start - [API overview](https://deference.si/docs/api-reference): Call models from your code with the OpenAI and Anthropic formats you already use. - [Authentication](https://deference.si/docs/api-reference/authentication): Send your API key as a bearer token or in x-api-key. - [SDKs and clients](https://deference.si/docs/api-reference/sdks-and-clients): Use the OpenAI and Anthropic SDKs, the Vercel AI SDK, LiteLLM and LangChain by changing the base URL. - [OpenAI compatibility](https://deference.si/docs/api-reference/openai-compatibility): Use any OpenAI SDK or tool by changing the base URL and the key. - [Anthropic compatibility](https://deference.si/docs/api-reference/anthropic-compatibility): Use Claude Code and Anthropic SDKs with the Messages API. ### Endpoints - [Chat completions](https://deference.si/docs/api-reference/chat-completions): Create a model response from a list of messages, with optional streaming, images and tools. - [Messages](https://deference.si/docs/api-reference/messages): Create a message with the Anthropic Messages API. This is the endpoint Claude Code uses. - [Responses](https://deference.si/docs/api-reference/responses): Create a response with the OpenAI Responses API. This is the endpoint Codex uses. - [Embeddings](https://deference.si/docs/api-reference/embeddings): Turn text into vectors with an embedding model. - [Images](https://deference.si/docs/api-reference/images): Generate images from a prompt with an image model. - [Models](https://deference.si/docs/api-reference/models): List the catalog or fetch one model, with prices and capabilities. - [Key](https://deference.si/docs/api-reference/key): Read the calling key's limit and spend, and the account's available credit and free credit. ### Guides - [Streaming](https://deference.si/docs/api-reference/streaming): Receive tokens as they are generated on chat completions, messages and responses. - [Tool calling](https://deference.si/docs/api-reference/tool-calling): Let a model call functions you define, then send the results back. - [Structured outputs](https://deference.si/docs/api-reference/structured-outputs): Constrain a response to a JSON schema. - [Prompt caching](https://deference.si/docs/api-reference/prompt-caching): Reuse a long prompt prefix at the cache read price. ### Reference - [Streaming events](https://deference.si/docs/api-reference/streaming-events): The event sequence each endpoint sends when stream is true. - [Errors](https://deference.si/docs/api-reference/errors): The error envelopes, status codes and error codes the API returns, and how to fix each. - [Rate limits](https://deference.si/docs/api-reference/rate-limits): Requests per minute, requests in flight, credit and the limits that apply to images. - [Docs for agents](https://deference.si/docs/api-reference/docs-for-agents): Every page has a Markdown twin, and llms.txt indexes them, so an agent can read these docs without scraping HTML. ## Legal - [Terms of service](https://deference.si/terms.md): The terms for using the Deference dashboard, API and docs. - [Privacy policy](https://deference.si/privacy.md): What we keep about you, where your prompts go, and your choices. - [Cookies and storage](https://deference.si/cookies.md): Everything Deference stores in your browser, and why. - [Acceptable use](https://deference.si/acceptable-use.md): What you cannot do with Deference, and what happens if you do.