Explore the documentation
Get started
Local quickstart
Run the NeuraMesh frontend and Node API locally, check prototype health, inspect reference GPU markets and stage an order without moving real funds.
End-to-end pipeline
Follow the proposed NeuraMesh pipeline from provider approval and capacity discovery to contracts, workload execution, evidence, disputes and Sui settlement.
Protocol & delivery
System architecture
Understand NeuraMesh service boundaries: discovery, Sui claim state, execution orchestration, Walrus evidence, verification and the current prototype gap.
Capacity, claims & reservations
Learn the proposed NeuraMesh object model: bounded GPU capacity pools, transferable compute claims, workload reservations and supply-preserving invariants.
Provider onboarding
Understand proposed NeuraMesh provider onboarding, hardware and capacity verification, metering obligations, collateral, failure procedures and pilot readiness.
Evidence & settlement
Follow NeuraMesh delivery evidence from reservation-bound receipts and Walrus references to independent verification, disputes, Sui settlement and failure handling.
Markets & economics
Spot, forwards & hedges
Compare the proposed NeuraMesh spot, monthly forward and perpetual-style markets, including GPU-hour terms, matching, liquidity and delivery boundaries.
Pricing & reference index
Learn how NeuraMesh proposes to construct a compute-price index using approved quotes, executed activity, external benchmarks and data-quality controls.
Margin & risk controls
Read the proposed NeuraMesh isolated and cross-margin model, funding controls, liquidation sequence, insurance backstop and production risk prerequisites.
Payments & reconciliation
Understand proposed NeuraMesh card, bank, USDC and NMESH payment flows, the separation of fiat and trading balances, and required reconciliation controls.
NMESH & revenue model
Read the proposed NMESH compute discount, governance, staking and rebates, alongside NeuraMesh revenue streams, treasury obligations and economic limits.
Build & reference
Prototype API reference
Inspect all six implemented NeuraMesh prototype endpoints, request examples, fixture responses, staging behavior, error codes and production limitations.
Roadmap & future prospects
Explore the investor-brief roadmap from controlled testnet to a brokered GPU pilot and scaled market, including engineering gates, metrics and future opportunities.
Glossary & source notes
Find plain-language definitions of GPU-hours, compute claims, reservations, metering, margin and settlement, plus source and implementation-status notes.
What does NeuraMesh do?#
Compute procurement needs more than a GPU name and an hourly price. A buyer needs to know where the hardware runs, when it is available, how much capacity is committed, what counts as delivery and what happens if delivery fails. NeuraMesh proposes a common contract language for these terms, with a market for discovery and an evidence trail for settlement.
The base commercial unit is one GPU-hour. It is a quantity of accelerator time under specified conditions, not ownership of a physical GPU. Two offers with the same GPU-hour quantity can differ in memory, topology, software compatibility, region and delivery obligations. Standardization makes these differences explicit; it does not make every offer interchangeable.
Choose your path through the docs#
Start with the pipeline if you want the complete operating model. Start with the local quickstart if you want to run the current implementation. Providers should read the onboarding and evidence guides before interpreting a capacity listing as an approved offer.
Local quickstart
Run the NeuraMesh frontend and Node API locally, check prototype health, inspect reference GPU markets and stage an order without moving real funds.
End-to-end pipeline
Follow the proposed NeuraMesh pipeline from provider approval and capacity discovery to contracts, workload execution, evidence, disputes and Sui settlement.
Provider onboarding
Understand proposed NeuraMesh provider onboarding, hardware and capacity verification, metering obligations, collateral, failure procedures and pilot readiness.
Prototype API reference
Inspect all six implemented NeuraMesh prototype endpoints, request examples, fixture responses, staging behavior, error codes and production limitations.
What is implemented today?#
The React/Vite site includes a finance workspace and reference markets. A dependency-light Node server returns fixed market and portfolio fixtures and stages order and settlement records in process memory. A successful request to that server does not execute a market trade, reserve a GPU or transfer funds.
The repository also contains Move and client scaffolds. These are not evidence of a configured, audited deployment. The client includes placeholder package IDs, and the existing Move marketplace models individual ComputeJob objects rather than the complete capacity-pool architecture. Treat the brief as the proposed destination and the inspected code as the description of current behavior.
| Area | Current state | Production requirement |
|---|---|---|
| Market data | Reference fixtures | Verified supply, fresh quotes and an operational index |
| Orders | In-memory staging | Durable state, authentication, risk checks and matching |
| Settlement | Prototype records and separate code scaffolds | Signed transactions, evidence review and dispute operations |
| Payments | No live movement | Payment partners, identity checks and reconciliation |
How to read status and examples#
Prototype behavior describes code that is present in this repository, not a promise about a hosted production service. Proposed design describes the investor brief or architecture document. A production gate is work that must be completed and verified before an operational claim can be made.
Prices, balances and dates shown in examples are fixtures or labeled illustrations. No launch date, provider availability, token return or support commitment is implied. The roadmap is sequenced by evidence and readiness, not by announced calendar deadlines.
Common questions
Is NeuraMesh a GPU cloud provider?
The proposed role is a coordination, market and settlement layer between buyers and approved compute providers. The current prototype does not itself establish verified live GPU inventory.
Where should a developer begin?
Run the local quickstart, inspect the six implemented API endpoints, then read the architecture and settlement guides before designing a production integration.
Sources & implementation status
Based on the NeuraMesh investor brief and the project sources below. Source review: . Design explanations are not a deployment audit, legal opinion or commitment to launch.
- Investor brief
- Marketplace architecture
- Node API prototype source
- Move marketplace source
- Client scaffold source
Read our source and status conventions. Suggest a correction.