Sell or Buy AI Models & Pipelines on ai.market
AIM-Node runs on your infrastructure and connects to ai.market's distributed peer-to-peer network. Sellers wrap models as MCP tools and publish. Buyers discover, purchase, and connect directly with no centralized proxy, no model uploads, and no middleman in the payload path.
For Sellers
Publish models without giving up control
Step 1
Install via pip
Install AIM-Node on your server with Python. It runs alongside your model or pipeline, with no uploads and no proxy layer.
Step 2
Wrap Model as MCP Tool
Expose your model or pipeline through AIM-Node's MCP interface. Define schemas, pricing, and clear descriptions for buyers and agents.
Step 3
Publish Metadata
Register with ai.market so discovery metadata, schemas, and descriptions are searchable while your weights and code stay private.
Step 4
Earn per-call
Buyers connect peer-to-peer to your AIM-Node. Usage is metered and settled through ai.market while you control pricing.
For Buyers
Buy access, then connect directly
Step 1
Browse
Find models and pipelines on ai.market by capability, schema, price, or quality score. AI agents can discover the same tools programmatically.
Step 2
Install consumer mode
Run AIM-Node on your own infrastructure in consumer mode so your application has a local control point for auth and connection management.
Step 3
Purchase token
Purchase access on ai.market and receive a cryptographically signed, time-limited delivery token scoped to that transaction.
Step 4
Connect P2P
Present the token and connect directly to the seller's AIM-Node over an encrypted peer-to-peer channel for live inference or pipeline execution.
Peer-to-Peer by Design
No Uploads. No Proxies. No Middleman.
What ai.market handles
Discovery, authentication, billing, delivery tokens, dispute resolution, and quality scoring. The platform coordinates the transaction and the money flow.
What ai.market never touches
Model weights, training data, inference inputs, inference outputs, and raw payloads. Those stay on the endpoints and move only between nodes.
How delivery works
After purchase, ai.market issues a signed delivery token. The buyer's AIM-Node presents it to the seller's AIM-Node, which verifies the token and opens the encrypted channel.
Why this matters
Direct delivery preserves performance, privacy, and control. Sellers keep custody of their stack, buyers get low-latency execution, and the delivery path has no central bottleneck.
Seller Node
Seller's Infrastructure
AIM-Node provider
Your model or pipeline
ai.market
Coordination Layer
Discovery + billing
Delivery token
Buyer Node
Buyer's Infrastructure
AIM-Node consumer
Your application
Metadata only
Seller Node ↔ ai.market
Metadata + token
ai.market ↔ Buyer Node
Direct Delivery Path
Encrypted P2P channel for inference data and results
Get Started with AIM-Node
Start with `pip` for Python-native installs or use Docker for containerized deployments. Full setup docs live at `get.ai.market`.
Recommended
pip install
pip install aim-nodeFor Python environments on any OS. Requires Python 3.10+.
Container
Docker
docker pull ghcr.io/aidotmarket/aim-node:latestdocker run -p 8400:8400 -p 8401:8401 ghcr.io/aidotmarket/aim-node:latestFor containerized deployments. Requires Docker.
System Requirements
Python 3.10+ or Docker
Use `pip install aim-node` for Python environments or Docker for containerized deployments.
4 GB RAM minimum
Enough for lightweight serving and local consumer-mode connectivity.
Network access
Outbound access to the ai.market API and peer-to-peer connectivity for delivery sessions.