<!-- TokenOS docs · Getting started · Markdown mirror of https://tokenos.ai/docs/whats-new · HTML: https://tokenos.ai/docs/whats-new · index: https://tokenos.ai/llms.txt -->

# What's new in TokenOS v5.1

v5.1 turns TokenOS from a build tool into an agent platform, and hardens the build tool itself: previews are verified in a real browser, images are real, wallets work inside the sandbox, and every change is reviewable.

October 2026P2P AI — peer-to-peer inference

TokenOS now runs a peer-to-peer inference market alongside the DeAI Compute Network. Anyone with spare API quota or a local model can sell inference; anyone can buy it with credits or pay per request over x402 with USDC on Solana. Start at [/p2p](https://tokenos.ai/p2p), or install the CLI with `npm i -g tokenos-p2p`.

| Feature | What it does | Docs |
| --- | --- | --- |
| P2P AI Inference Market | A peer-to-peer market for LLM inference at /p2p. Sellers publish OpenAI-compatible offers signed with a Solana identity; buyers pay with credits (USDC or $TOS) for metered usage only. The live catalog shows every model with its savings against list price. | [Read](https://tokenos.ai/docs/p2p-ai) |
| Sell inference without a GPU | Any API key (OpenAI, Anthropic, Google, DeepSeek…) or a local Ollama / vLLM model becomes a provider in ten minutes with the tokenos-p2p seller node. Payouts are USDC on Solana; running a node does not guarantee traffic or income. | [Read](https://tokenos.ai/docs/p2p-sell) |
| Buyer node & keys | Create tos_ keys with spend limits and expiry, run the local buyer node (OpenAI-compatible on:8377) or point any OpenAI SDK at it. Cheapest-then-trust routing, per-session budgets, signed receipts and dispute ceilings. | [Read](https://tokenos.ai/docs/p2p-buy) |
| Hosted gateway · x402 pay-per-request | No node, no account: POST /api/p2p/v1/chat/completions, /messages or /responses answers an x402 v2 challenge priced per request. Pay with USDC on Solana (tx: or PAYMENT-SIGNATURE) and the gateway routes to the cheapest trusted peer. | [Read](https://tokenos.ai/docs/p2p-buy) |
| Trust scores & leaderboard | Every peer is scored 0–100 from settled history, weekly usage share and verified identity (GitHub portfolio or domain). Wash-flagged peers drop to zero; the leaderboard ranks sellers by settled USD and unique buyers. | [Read](https://tokenos.ai/docs/p2p-ai) |
| Price history & alerts | 30-day best-price sparklines per model, a live settlement ticker, and buyer price alerts that e-mail you when a model falls below your target. | [Read](https://tokenos.ai/docs/p2p-buy) |
| Seller digest & uptime tooling | Weekly settled-USD digest for providers, systemd / pm2 recipes, upstream presets (flex tiers, quotas) and a clean-shutdown protocol that settles open sessions before exit. | [Read](https://tokenos.ai/docs/p2p-uptime) |

v5.1 liveHeadline changes

| Feature | What changed | Docs |
| --- | --- | --- |
| Autopilots | Persistent agents with a schedule (manual · interval ≥ 15 min · daily · weekly), a tool belt, skills and attached MCP servers. Up to 12 per wallet; every run is recorded step by step and gated tools pause for approval. | [Read](https://tokenos.ai/docs/autopilots) |
| Autopilot Computer | A white-labeled isolated cloud sandbox per run (Python · JavaScript · Bash). Billed per second of sandbox time at compute cost + 20% in credits, settled when the run ends. | [Read](https://tokenos.ai/docs/autopilots) |
| Approvals Inbox | High-impact actions (agent calls, e-mail, on-chain writes) wait in a web inbox at /autopilots. Approve, reject or read the run report — no terminal required. | [Read](https://tokenos.ai/docs/autopilots) |
| MCP server · A2A agent cards | TokenOS is a stateless streamable-HTTP MCP server (protocol 2026-07-28) and an MCP client. Every marketplace agent is an MCP tool and a signed A2A agent card; paid calls settle over x402. | [Read](https://tokenos.ai/docs/mcp-a2a) |
| ERC-8004 identities | Identity and reputation registries on Robinhood Chain (4663) and Base (8453). Creators register their agents on chain with their own wallet; anyone can read the trust profile. | [Read](https://tokenos.ai/docs/erc8004) |
| Jev · System One | A fail-open risk and acceptance engine that classifies build requests, routes model tiers, gates Autopilot approvals and is the last stage of harness verification. | [Read](https://tokenos.ai/docs/jev) |
| Build harness & Best-of-N | Plan → sprint → evaluate loops with self-healing, or run up to three models in parallel and apply the winning snapshot from the Agents window (⌘J). | [Read](https://tokenos.ai/docs/build-pipeline) |
| Build quality gate | Every preview is mounted in a headless Chromium before the build is marked complete. Runtime errors are fed back to the model for up to two repair rounds; the Render badge shows the outcome. | [Read](https://tokenos.ai/docs/quality-gate) |
| Brand art generation | Hero, feature and logo imagery referenced under /media is generated for real (Gemini image model for rasters, procedural SVG for logos) and materialised into the preview and any hosted copy. | [Read](https://tokenos.ai/docs/quality-gate) |
| Preview sandbox + wallet bridge | Previews run in an opaque-origin sandbox (SEC-003). Wallet access is relayed through a consent-gated postMessage bridge to the wallet you connected in the IDE. | [Read](https://tokenos.ai/docs/preview-sandbox) |
| Inline diff review | Iterate changes apply instantly and are listed hunk by hunk with Accept / Reject / Undo, per-file accept / revert and Accept all. | [Read](https://tokenos.ai/docs/iterate) |
| Sessions rail | The explorer lists your harness, Best-of-N and Autopilot runs live, with a shortcut to pending approvals. | [Read](https://tokenos.ai/docs/workspace) |
| Dynamic credit pricing | Packages are priced in USD and re-priced twice a day against live SOL / USDC / $TOS prices. Paying with $TOS is always 20% cheaper. | [Read](https://tokenos.ai/docs/credits) |
| One signature per 30 minutes | Authenticated reads use a wallet session token minted from one signature; only writes (build, iterate, deploy, launch, approve) sign per action. | [Read](https://tokenos.ai/docs/security) |
| Unified fee destinations | Every platform fee destination and the platform fee rate (300 bps) are configured server-side and surfaced to payers only through the x402 quote. Nothing is hardcoded in the client. | [Read](https://tokenos.ai/docs/x402) |

### Carried forward from v4

v4 delivered the rewritten generation pipeline (two parallel passes per page, truncation retry, cross-provider fallback), the DeAI Compute Network, the hardened on-chain credit system (purchase verified on chain, replay-protected, admin-only deductions), public contract pages, the leaderboard and fork, and token launches on Pump.fun. All of it ships in v5.1.

### Behaviour changes worth knowing

- A build is no longer *completed* when files exist — it completes when the preview bundles **and** the render check finishes. Expect the status to stay *processing* a little longer.
- Generated apps never contain a keyed RPC URL. Solana mainnet apps call the allow-listed proxy at `/api/solana-rpc`; confirmation uses HTTP polling, not WebSockets.
- Pages receive `wallet` as an *address string*; the runtime (`lib/web3.js`, `hooks/useWallet.js`, `lib/walletBridge.js`, `config.js`) is protected from the JSX sanitizers and must not be edited by hand inside the IDE.
- Legacy model ids (for example `claude-sonnet-4-20250514`) are remapped to their current successors and hidden from pickers.
- Credit cost for a new build is `max(5, round(7.5 × multiplier × effort)) × prompt tier` — see [Credits](https://tokenos.ai/docs/credits).

---

*Getting started · TokenOS documentation. Canonical page: https://tokenos.ai/docs/whats-new. All docs as Markdown: https://tokenos.ai/llms.txt. Payments on TokenOS are for metered usage only; nothing on this page is investment advice or a promise of income.*
