documentation

Quick start

Connect to MeshMoE in 3 steps — OpenAI-compatible API, zero migration cost.

① Sign up and create an API key

You get 500 free credits on sign-up. After logging in, create a key on the API Keys page (format moe-xxxxxxxx).

② Make your first request

# curl example — mesh-pro intelligent routing
$ curl https://api.meshmoe.com/v1/chat/completions \
  -H "Authorization: Bearer moe-your-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "mesh-pro",
    "messages": [{"role": "user", "content": "hi, introduce yourself"}]
  }'
  

# Python example (openai SDK)
>>> from openai import OpenAI

client = OpenAI(
    api_key="moe-your-key",
    base_url="https://api.meshmoe.com/v1"
)

response = client.chat.completions.create(
    model="mesh-pro",  # mesh-free / mesh-pro / mesh-max / mesh-research
    messages=[{"role": "user", "content": "write quicksort in Python"}]
)
print(response.choices[0].message.content)
  

③ Check your usage

Token consumption, credit deductions, and model-selection logs for every request are shown in real time on the usage page.

The four MeshMoE tiers

model	role	price ($/M)	notes
`mesh-free`	free	$0	daily free quota, node + free API
`mesh-pro`	workhorse	$0.30 / $0.60	intelligent routing, 80% of traffic
`mesh-max`	flagship	$0.90 / $2.00	single flagship model, for hard problems
`mesh-research`	deep research	$3.00 / $8.00	MoA multi-model synthesis, major decisions

Desktop client

Download the MeshMoE desktop client (coding agent + node contribution + four-tier AI, all in one): download page →

# desktop client config (~/.meshmoe/config.toml)
default_provider = "meshmoe"

[[providers]]
name = "meshmoe"
base_url = "https://api.meshmoe.com/v1"
api_key = "moe-your-key"
models = ["mesh-free", "mesh-pro", "mesh-max", "mesh-research"]
  

Compatibility

The MeshMoE API is fully OpenAI-compatible.
Just change base_url to https://api.meshmoe.com/v1 and model to mesh-pro (or another tier) — and you're good to go.

Third-party tool integration

MeshMoE works with any tool that speaks the OpenAI API format. Configuration for the most popular tools:

Claude Code

# set environment variables (MeshMoE is Anthropic-format compatible)
$ export ANTHROPIC_BASE_URL=https://api.meshmoe.com/v1
$ export ANTHROPIC_API_KEY=moe-your-key

# or write to ~/.claude/.env
ANTHROPIC_BASE_URL=https://api.meshmoe.com/v1
ANTHROPIC_API_KEY=moe-your-key
  

Cursor

Settings → Models → OpenAI API:
1. Check "OpenAI API Key"
2. Enter API Key: moe-your-key
3. Enter Base URL: https://api.meshmoe.com/v1
4. Model name: mesh-pro

Continue (VS Code)

# ~/.continue/config.json
{
  "models": [{
    "title": "MeshMoE Pro",
    "provider": "openai",
    "model": "mesh-pro",
    "apiBase": "https://api.meshmoe.com/v1",
    "apiKey": "moe-your-key"
  }]
}
  

OpenAI SDK (Python / Node.js)

# Python
>>> from openai import OpenAI
client = OpenAI(api_key="moe-your-key", base_url="https://api.meshmoe.com/v1")

# Node.js
$ import OpenAI from "openai"
const client = new OpenAI({ apiKey: "moe-your-key", baseURL: "https://api.meshmoe.com/v1" })
  

LangChain

>>> from langchain_openai import ChatOpenAI
llm = ChatOpenAI(
    model="mesh-pro",
    api_key="moe-your-key",
    base_url="https://api.meshmoe.com/v1"
)
  

Hermes / OpenClaw

# .env config
AI_API_KEY=moe-your-key
AI_BASE_URL=https://api.meshmoe.com/v1
AI_MODEL=mesh-pro
  

curl / any HTTP client

$ curl https://api.meshmoe.com/v1/chat/completions \
  -H "Authorization: Bearer moe-your-key" \
  -H "Content-Type: application/json" \
  -d '{"model":"mesh-pro","messages":[{"role":"user","content":"hi"}]}'
  

Every tool just needs 3 values changed: base_url → api.meshmoe.com/v1, api_key → moe-xxx, model → mesh-pro (or free/max/research).

Get started free →