Feature · Providers
Manage every LLM provider and API key in one place
Stop pasting the same API key into five agent configs. Add each provider to AutoJev once, test it, set model prices, and every connected agent can use it through the local gateway.
- Presets for OpenAI, DeepSeek, OpenRouter and Ollama
- Any OpenAI- or Anthropic-compatible endpoint
- Import from CC Switch and Termany
- Test, enable or disable each provider
Providers
OpenRouter
https://openrouter.ai/api
DeepSeek
https://api.deepseek.com
OpenAI
https://api.openai.com/v1
Ollama
http://127.0.0.1:11434
Bring your own keys — once
Each coding agent wants its own copy of your provider settings. Rotate a key or try a new provider and you're editing config files across the machine again.
In AutoJev, providers live in one list. Agents never see your real keys: they talk to the local gateway at http://127.0.0.1:9527/v1, and AutoJev forwards each request directly from your machine to the provider. There's no AutoJev account, no cloud relay and no markup on tokens.
Built-in presets
Presets fill in the endpoint for you. Two generic presets cover everything else.
Use the compatible presets for any OpenAI-compatible endpoint such as Kimi (Moonshot), GLM (Zhipu), Qwen (DashScope), SiliconFlow, vLLM or LM Studio, or any Anthropic-compatible endpoint.
Models carry the details routing needs
Under each provider, add the models you want to use. Their metadata drives routing, protocol conversion and cost estimates.
API type
chat_completions, responses or messages — the API the upstream actually offers. AutoJev converts when an agent speaks a different one.
Prices
Input, output and cache prices per 1M tokens. Used for estimated cost in logs and for the lower-cost routing preference.
Tier and capabilities
Tier plus tools, vision and reasoning flags and the context window. Routing hints, not hard limits — the upstream decides what it accepts.
Import what you already have
Already set up providers in another tool? Import them instead of retyping.
- 1
CC Switch
Reads Claude Code and Codex providers from ~/.cc-switch/cc-switch.db, opened read-only.
- 2
Termany
Reads providers from ~/.termany/termany.db, also read-only.
- 3
De-duplicated by endpoint
Providers whose endpoint already exists in AutoJev are skipped, so importing twice won't create copies.
Test before you route
- Run a provider test to confirm the URL, key and test model respond
- Replies that contain only reasoning still count as a pass — the key, URL and model work
- Failures show the HTTP status and the provider's error, with your key redacted
- Tests use the same outbound network settings as the gateway: direct, system proxy or a custom proxy
- Disable a provider to take all its models out of rotation without deleting anything
How agents see your providers
Every enabled model becomes available on the gateway as <provider>/<model_id>, and can be added to any route such as autojev/fast. GET /v1/models lists the models and routes an agent can use, so clients with a model picker discover them automatically.
Swapping providers becomes a change in AutoJev, not in your agents. Move a route from one provider to another, add a backup, or disable a provider that is misbehaving — connected agents keep using the same route ID and never need their config touched again.
Where your keys are stored
Provider configuration and API keys are saved locally in a SQLite database at ~/.autojev/autojev.db. Keys are currently stored unencrypted, so protect that file like any other credential on your machine and don't commit it or its backups to a repository.
Frequently asked questions
Which LLM providers does AutoJev support?
OpenAI, DeepSeek, OpenRouter and Ollama have presets. Any other OpenAI-compatible or Anthropic-compatible endpoint works through the compatible presets with a custom URL.
Can I use a self-hosted model server like vLLM or LM Studio?
Yes, as long as it exposes an OpenAI-compatible or Anthropic-compatible API. Add it with the OpenAI Compatible or Anthropic Compatible preset and your server's URL.
Are my API keys encrypted?
No. Keys are stored locally in ~/.autojev/autojev.db and are currently unencrypted. They never leave your machine except in requests to the provider they belong to, and they're not written to request logs.
How do I import providers from CC Switch?
Use the import option on the Providers page and choose CC Switch. AutoJev reads ~/.cc-switch/cc-switch.db without modifying it and skips any endpoint you already have.
Does AutoJev charge for tokens?
No. Requests go straight from your machine to your providers with your own keys. AutoJev is free and adds no markup.
Keep exploring
OpenAI-compatible APIsKimi, Qwen, SiliconFlow, MiniMax, vLLM, LM Studio — any OpenAI-compatible endpoint, available to all your agents.IntegrationsOllamaUse local Ollama models in any coding agent, prefer them in routing, and fail over to cloud models when needed.CompareAutoJev vs CC SwitchPer-request model routing and speed tests versus one-click provider switching with MCP, Skills and prompt management.FeaturesSmart model routingExpose routes like autojev/fast to every agent. Load-balance models or let Jev pick by cost, quality, speed or balance.FeaturesDebug consoleTest models and routes in conversation or cURL mode and inspect the selected model, timing, tokens, cost and raw JSON.Updated 2026-10-14
Give every agent the right model.
Install AutoJev, add the providers you already pay for, and connect your agents in one click. Everything stays on your machine.
Free & open source · macOS · Windows · Linux