Cheapest
The call goes to the lowest price among sellers who are online and under their rate limit.
policy: "cheapest"4663 // rt-01
P2P inference marketPick who runsyour model
Sellers list a price for each model they host. You choose one, keep the tools you already use, and pay per call up to a daily limit you set.
- Models
- 5
- Sellers online
- …
- Cheapest chat-70b / 1M
- …
- Market
- open board →
One endpoint on your machine
Your OpenAI-style tools talk to a router on localhost. The router decides which seller gets the call.
A price for every model
Sellers set their own rate per 1M tokens. You see it before anything is sent.
Paid from escrow
You top up once. Each finished call is paid from that balance and comes with a signed receipt.
03Route
How it works
Three waysto route a call
Every model gets its own rule. Most people leave chat on cheapest and pin the one seller they trust for code.
Fastest
Picks the seller with the best response time over the last few minutes. Usually costs a bit more.
policy: "fastest"Pinned
Always the same seller. If they go offline the call fails, it doesn't quietly move somewhere else.
policy: "pinned"04Control
Advantage
Price before you send
The rate per 1M tokens is on the seller's listing. Nothing gets billed at a number you didn't see.
A hard daily limit
Set a USD cap. Once you hit it, paid calls stop until 00:00 UTC. There's no credit and no overdraft.
Drop a seller anytime
Routes are set per model. Remove a seller and your next call goes to someone else on the same endpoint.
05Network
Who it'sfor
Buy inference for the tools you use, or rent out a GPU that sits idle half the day.
AI 01
Buyers
Keep your chat app, editor plugin or script. Change one base URL and compare prices per seller.
AI 02
Sellers
Run a model on your own hardware, set the price and the limits. You earn when people route to you.
AI 03
Agents
Give an agent a budget and a routing rule. It spends from escrow and stops at the cap.
AI 04
Developers
Same request and response shape as the OpenAI API. Your code stays as it is.
06Stack
Integration
Swap onebase URL
Anything that speaks /v1/chat/completions works. Point it at the local router and pick the model by name.
- Chat apps with a custom endpoint field
- Editor plugins like Continue or Cline
- OpenAI SDKs for Python and Node
- Plain curl
The local router is still being built. Until it ships, the playground runs the same flow against simulated sellers.
07Token
Launch progress
Launchingon Pons
Not launched yet
There's no LANEX token yet and no address. When it launches it starts on a Pons bonding curve priced in ETH. Once 4.2 ETH is in the curve it graduates into a Uniswap V4 pool. Until then the chart on the right is another Pons launch, shown as a sample.
64% to graduation (2.69 / 4.2 ETH) · samplegraduation 4.2 ETH
| Recent Pons launches | sample |
|---|
Price curve sample0.00000001093 ETH
—Merch drop 01
Caps andtees
Six pieces in navy, olive and black, with the mark stitched on the front. Price and drop date aren't set. Drag the model to turn it.
Dad cap · navy
Washed cotton
Unstructured six-panel cap with a metal buckle. Tees are heavyweight, boxy fit.
Mark on the front
Embroidered on the caps, screen printed on the tees. Small, left of center.
No price yet
Join the list below and you'll hear about it before it goes on sale.
Early accesslist
Leave an email for the merch drop and launch news. You can take it off the list whenever you want.
08Start
Get started
Prices onthe board
Try the playground
Pick a model and a rule, send a prompt and look at who answered and what it cost.
Open playground →Read the board
Every live offer with price, speed and load. Sort by whatever matters to you.
Market board →List your GPU
Set a price per 1M tokens, a rate limit and a daily token limit. Change them anytime.
Sell capacity →