App setup guides
Tested, step-by-step setup for 24 apps. Each guide lists the exact values to paste, the steps in the app, and fixes for the errors people run into. Every guide was checked against the app's own source or documentation, and Claude Code and Codex were also run end to end against the live API.
All guides need the same three things: a FreeTheAI account, an API key, and today's check-in.
| You need | Value |
|---|---|
| Base URL (most apps) | https://api.freetheai.org/v1 |
| Full chat URL (JanitorAI, TypingMind) | https://api.freetheai.org/v1/chat/completions |
| Anthropic base URL (Claude Code) | https://api.freetheai.org |
| Model | An fta/ ID from freetheai.org/models, such as fta/zai/glm-5.3 |
Your app is not listed? Any app with a custom OpenAI-compatible base URL works: use the base URL above, your key, and a model ID.
Roleplay apps
| App | How it connects |
|---|---|
| JanitorAI | Proxy settings: the full chat completions URL, your key, and a model ID, then a page refresh. |
| SillyTavern | Chat Completion with the Custom (OpenAI-compatible) source. Lists every model your key can use. |
| Chub AI | An OpenAI reverse proxy in Secrets, then the model in Configuration. |
| RisuAI | Custom API model (OpenAI Compatible). On the web, turn on Force Plain Fetch. |
| Tavo | A Custom (OpenAI Protocol) connection on Android and iPhone. |
For roleplay, start with fta/zai/glm-5.3. The fta/dog/ and fta/lion/ models refuse roleplay. More in Models and IDs.
Chat apps
| App | How it connects |
|---|---|
| Open WebUI | An OpenAI API connection that lists every model your key can use. |
| LibreChat | A custom endpoint in librechat.yaml, with the key in .env. |
| LobeHub (LobeChat) | A custom provider with the OpenAI request format. |
| Chatbox | OpenAI API Compatible mode, on desktop, mobile, and the web. |
| Cherry Studio | An OpenAI-type provider; enter the root address, Cherry adds /v1. |
| NextChat | A custom OpenAI endpoint and a custom model list. |
| TypingMind | Custom models with the full chat completions URL and an Authorization header. |
| AnythingLLM | The Generic OpenAI provider. |
| Page Assist | An OpenAI compatible provider in the browser extension. |
Coding tools
| Tool | How it connects |
|---|---|
| Claude Code | ANTHROPIC_BASE_URL and a tested settings.json, through Anthropic Messages. |
| Codex CLI | A custom model provider in config.toml, through the Responses API. |
| OpenCode | /connect for the key and a provider in opencode.json. |
| Kilo Code | An OpenAI Compatible custom provider in VS Code and the CLI. |
| Cline | The OpenAI Compatible API provider in VS Code. |
| Continue | A model entry in config.yaml with the openai provider. |
| Aider | OPENAI_API_BASE, OPENAI_API_KEY, and the openai/ model prefix. |
| Zed | An OpenAI compatible provider for the agent panel. |
| VS Code Copilot Chat | The Custom Endpoint provider (bring your own key), even without a Copilot plan. |
| Hermes Agent | A custom endpoint with hermes model or config.yaml. |
Coding tools need tool calls. Start with fta/zai/glm-5.3, and turn on Tool-capable models only on the key you use for agents (key settings). Free accounts run one request at a time; most coding tools retry the extra requests on their own.