TokelyConnect AI coding agents to Claude, GPT, Kimi K2.7 Code, DeepSeek and Gemini through Tokely's OpenAI-compatible endpoint. Copy-paste settings for each tool.
By TokelyReviewed Coding agents that accept an OpenAI-compatible provider work with Tokely: set the base URL to https://api.tokely.me/v1, paste your Tokely key and choose a model ID. Claude Code uses the native Anthropic endpoint instead. One balance covers every model, with prompt caching and per-key budgets.
# Cline, Roo Code, Kilo Code (VS Code): Settings -> API Provider
API Provider: OpenAI Compatible
Base URL: https://api.tokely.me/v1
API Key: sk-your-tokely-key
Model ID: kimi-k2.7-code (or claude-sonnet-5, deepseek-v4-pro-0813, gpt-5.4)
# OpenCode: opencode.json
{
"provider": {
"tokely": {
"npm": "@ai-sdk/openai-compatible",
"options": { "baseURL": "https://api.tokely.me/v1", "apiKey": "{env:TOKELY_API_KEY}" },
"models": { "kimi-k2.7-code": {}, "claude-sonnet-5": {} }
}
}
}
# Continue: config.yaml
models:
- name: Kimi K2.7 Code
provider: openai
model: kimi-k2.7-code
apiBase: https://api.tokely.me/v1
apiKey: sk-your-tokely-key
# Claude Code: see /claude-code-api (native /v1/messages)Base URL https://api.tokely.me/v1. Create a key in the dashboard and keep it in an environment variable. Quickstart · Model IDs and endpoints
Use a strong model such as Claude Sonnet for planning and hard changes, and a low-cost model such as Kimi K2.7 Code or DeepSeek V4 Pro for routine edits and background tasks. Changing the model is a change of model ID in the agent's settings.
Agents send large, repeated context. Prompt caching bills the repeated prefix at the cache-read rate, and a dedicated API key per developer or agent with its own budget stops a runaway session before it is billed.
Editors with an override for the OpenAI base URL can point their chat models at Tokely. Some editor features still run on the editor's own backend and do not use your key; check the editor's documentation.
Yes, through the native /v1/messages endpoint. See the Claude Code guide for the environment variables.
Start with Claude Sonnet for quality and Kimi K2.7 Code for cost, then compare cost per accepted change on your own tasks.
Check the model, compare the rate, and connect with a Tokely key.