live network
Community compute for AI agents
Spare CPU/RAM/GPU and open-model LLM inference, shared not sold. Every number on this page is real and updates automatically -- nothing here is a mockup.
Add it to your agent
Same command in Claude Code or Gemini CLI -- get an API key first, then swap it in below.
claude mcp add --transport http omnigrid https://chanza.ai/mcp.php \
--header "Authorization: Bearer YOUR_API_KEY"gemini mcp add --transport http omnigrid https://chanza.ai/mcp.php \
--header "Authorization: Bearer YOUR_API_KEY"0
providers online
0.072
compute-hours donated
34 / 43
jobs completed
12 / 17
multimodal (image) jobs processed
upload → base64 in RAM → sent to model → discarded
Being shared for free, right now
Leaderboard
Bragging rights only -- credits aren't spendable, just recognition for what you've contributed.
1
mexmarv
63.1
2
marvin
36.9
Recent activity
The last few jobs completed on the network.
Text generation (qwen3-8b-m4)
13.16s compute ·
Text generation (qwen3-8b-m4)
15.22s compute ·
Text generation (qwen3-8b-m4)
8.35s compute ·
Text generation (qwen3-8b-m4)
9.49s compute ·
Text generation (qwen3-8b-m4)
16.50s compute ·
Vision-language (llava-7b-m4)
10.00s compute ·
Donate spare CPU/RAM/GPU, or point your Omnigent/Claude Code/Cursor agent at community-hosted open models -- both start with an account and an API key.
Get your API key →