Providers observed in the index
Specialized models deserve a path to market.
npi · 24h
volume-weighted, never estimated
For creators
Turn local models into market-ready services.
Fine-tune an open-weight model, serve it on your own infrastructure or through a compute provider, and earn whenever users call it.
settled job
index
job 4a12
✓
job 4a13
✕
job 4a14
✕
For users
Access expertise per request.
Discover expert models built for fields like biomedicine, law, finance, and science. Pay for the inference you need without owning infrastructure.
frame n
0x9f3c…
✓
fold offline · compare head
For providers
Earn from real model usage.
Serve your own model or work with a compute provider. Each finalized request pays the provider, funds the protocol, and can unlock additional incentives.
From expert model
to paid inference.
Creators publish specialized models. Users bring demand. Providers execute requests, and finalized work produces revenue, incentives, and a transparent market record.
Settlement gate
All contract conditions must pass before providers are paid and activity becomes reputation.
Hash-chained transcripts
Every lifecycle transition is a signed SJP-1 frame — replayable by any third party.
Creator economics
Finalized usage pays providers and can add protocol incentives without changing the buyer quote.
How it works
From personalized & local model to paid inference.
Each step maps to a contract call. The SJP-1 transcript records every transition as a signed, hash-chained frame replayable offline by any third party.
Market outcomes
A market that rewards real expertise.
Specialized models reach global demand. Creators and compute providers earn from usage. Finalized settlements keep the resulting market history transparent.
Settlement intelligence
Transparent pricing from real paid work
Every finalized job leaves a public price record: what was paid, how much work was delivered, and which model and provider served it. Pricing on OpenAlmond comes from real paid work.
Specialized supply
Expert models become accessible services
Researchers, labs, fine-tuners, and teams can turn local and open-weight models into paid services built for specific domains.
Flexible serving
Serve directly or delegate compute
Compute providers are admitted, bonded participants who accept and deliver jobs. Provider-level views show the distribution of settled prices across a participant’s history. They describe realized settlement activity, not capability claims.
Open demand
Users pay per inference
Users, applications, and agents can request specialized intelligence through a shared market and pay only for the work they ask the network to execute.
Creator economics
Real usage creates revenue
Each finalized request pays the serving provider. Protocol incentives can add a separate reward on top, giving specialized models a path from research artifact to recurring service.
Fee transparency
Payment and fees stay explicit
The buyer sees the quoted payment before execution. The protocol fee is explicit, and gross versus net settlement records keep its effect visible without rewriting the buyer price.
Market history
Real work builds market reputation
Finalized jobs create a transparent price index and verifiable records of volume, models, and providers. Empty markets stay empty.
FAQs