# Candor API

> Uncensored language models for agents and apps, by MCP or an OpenAI-compatible API.

1. Sign in at https://candorai.dev. New accounts get $0.47 of free credit to try the API, once per person.
2. When the free credit runs out, subscribe at https://candorai.dev/?panel=account: Monthly $19.99/month, includes 8,500 credits ($20.00 of API usage); or Plus $99/month, includes 42,500 credits ($100.00 of API usage).
3. Optional top-ups at the same page: $9.99 for 4,250 credits ($10.00 of API usage), subscribers only.
4. Create a key under Account > API keys (shown once). A key can be given a spending limit in USD (in total, per day or per month) and an end date, for example one key per project or per agent: it stops at its limit, or on its date, without touching the rest of the account. Both can be changed later on the same screen.

Add Candor to Claude Code in one command:

    claude mcp add --transport http candor https://candorai.dev/mcp --header "Authorization: Bearer $CANDOR_API_KEY"

Remote MCP endpoint: https://candorai.dev/mcp (Streamable HTTP, header "Authorization: Bearer sk-unc-..."). Tools: chat, list_models, candor_balance, list_characters, web_search.
OpenAI-compatible base URL: https://candorai.dev/v1. Models: `candor-fast`, `candor-classic`, `candor-smart`. Streaming (SSE), tool calling and usage reporting follow the OpenAI Chat Completions format. New accounts get $0.47 of free credit to try the API, once per person. After that, prepaid credits from a subscription.

## Docs

- [Quickstart and client configs](https://candorai.dev/docs): 60-second setup, Claude Code, Cursor, Claude Desktop, Codex, Windsurf, Python, JavaScript, LangChain, Vercel AI SDK, and a guide for Cline and Roo Code, Continue, aider, Open WebUI, LibreChat, SillyTavern, n8n, LiteLLM
- [Full reference for LLMs](https://candorai.dev/llms-full.txt): every endpoint, parameter, error, price and limit in one file
- [OpenAPI 3.1 spec](https://candorai.dev/openapi.json): machine-readable schema
- [Postman collection](https://candorai.dev/postman.json): every endpoint ready to try; set the collection variable CANDOR_API_KEY
- [Instructions for coding agents](https://candorai.dev/agents.md): the exact facts and the rules that avoid the usual mistakes, written for an assistant wiring Candor into a project

## Endpoints

- [POST /mcp](https://candorai.dev/llms-full.txt): remote MCP server (chat, list_models, candor_balance, list_characters, web_search)
- [POST /v1/chat/completions](https://candorai.dev/llms-full.txt): create a chat completion (stream or not)
- [GET /v1/models](https://candorai.dev/v1/models): list models with prices (no key needed)
- [GET /v1/balance](https://candorai.dev/llms-full.txt): remaining balance in USD, free credit included
- [GET /v1/usage](https://candorai.dev/llms-full.txt): per-request usage and cost
- [GET /v1/usage/summary](https://candorai.dev/llms-full.txt): what was spent by day, by model and by key
- [POST /v1/web/search](https://candorai.dev/llms-full.txt): search the web; or add "web_search": true to a chat request and the model answers from the results
- [GET /v1/characters](https://candorai.dev/llms-full.txt): the characters that can be played; add "character" to a chat request to talk to one