API Documentation
MeshTok is a 100% OpenAI-compatible gateway. If your code already calls OpenAI, you only need to change two things: the base URL and your API key.
1. Get your API key
Sign in to the console and create a token. Your key looks like sk-xxxx....
2. Base URL
OpenAI-compatible (recommended)
https://meshtok.com/v1 Use with OpenAI SDK, Cursor, Cline, Aider, and most tools. Covers all 175 models.
Anthropic-compatible
https://meshtok.com Use with Claude Code, Anthropic SDK, and tools that require the Anthropic Messages API format.
Most users only need the OpenAI-compatible endpoint. All endpoints follow the OpenAI convention: /v1/chat/completions, /v1/models, /v1/embeddings.
3. Python example
# pip install openai from openai import OpenAI client = OpenAI( base_url="https://meshtok.com/v1", api_key="sk-your-MeshTok-key", ) # any of 175 models -same code, just change the model name resp = client.chat.completions.create( model="deepseek/deepseek-v4-pro", messages=[{"role":"user","content":"Explain quantum computing in one sentence."}], ) print(resp.choices[0].message.content)
4. JavaScript example
// npm install openai import OpenAI from "openai"; const client = new OpenAI({ baseURL: "https://meshtok.com/v1", apiKey: "sk-your-MeshTok-key", }); const resp = await client.chat.completions.create({ model: "bigmodel/glm-5.2", messages: [{ role: "user", content: "Hello!" }], }); console.log(resp.choices[0].message.content);
5. Streaming
Add stream: true for Server-Sent Events streaming, identical to OpenAI's format.
curl https://meshtok.com/v1/chat/completions \ -H "Authorization: Bearer sk-your-MeshTok-key" \ -H "Content-Type: application/json" \ -d '{"model":"anthropic/claude-sonnet-5","messages":[{"role":"user","content":"Hi"}],"stream":true}'
6. List all models
curl https://meshtok.com/v1/models \
-H "Authorization: Bearer sk-your-MeshTok-key" Returns the full list of 175 available model IDs. Browse them visually on the models page.
Feature guides
Streaming responses →
Server-Sent Events in the OpenAI format. Python, JS, raw SSE parsing.
Tool & function calling →
Let 154 models invoke your functions. Full agentic loop, parallel calls.
Structured outputs →
Force valid JSON. json_object and strict json_schema modes.
Vision & image input →
Send images to 58 vision models. URL, base64, OCR, multi-image.
Model fallbacks →
Automatic failover across labs. Sequential and parallel patterns.
Coding agents & IDE →
Claude Code, Cursor, Cline, Aider, Continue, Codex setup.
Multimodal input →
Images, audio, video, PDF — 69 multimodal models.
Ready to build? Get your API key