> For the complete documentation index, see [llms.txt](https://docs.brainnetwork.app/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.brainnetwork.app/participate/run-a-browser-node.md). # Run a browser node Five minutes, nothing installed. ## Requirements A browser with WebGPU: current Chrome or Edge on Windows, macOS or Linux; Safari 18 or later on macOS; Chrome on Android. Firefox support is partial and depends on your version. A discrete or Apple Silicon GPU gives a higher benchmark score; integrated graphics work and score lower. ## Steps 1. Open [brainnetwork.app/earn](https://brainnetwork.app/earn). 2. **Detect.** The page reads what WebGPU exposes and shows each field with its source. Anything the browser does not expose says *unavailable*. VRAM always does. 3. **Benchmark.** Click to run the challenge. A WGSL kernel runs on your GPU for under a second against a server-seeded secret; the server recomputes secret blocks and scores you on its own clock. Your score is shown with a **LIVE** badge because it was measured in your tab and verified on the server. 4. **Join.** Your node appears in the topology, the join counter increments by exactly one, and work starts arriving within seconds. 5. **Connect a wallet.** Phantom, Solflare, Backpack, or an email through Privy. Sign one message; it moves no funds. Verified units from your tab are attributed to this wallet. Leave the tab open. Minimise it, move it to another desktop, do not close it. ## What you will see * **Jobs completed, verified compute, reputation** on the dashboard, all server-computed and badged LIVE. * **This epoch accrued**: your verified work so far against the current pool, pro-rated to elapsed time. It becomes a real allocation when the hour settles. * Your node's page at `/node/`: verification rate, reassignment rate, median latency, uptime, all from server records. ## Claiming At two minutes past each hour the previous hour settles. On [/rewards](https://brainnetwork.app/rewards), connect the same wallet and claim. You sign a message binding the amount, a nonce and an expiry; the server sends SOL from the payout wallet and shows you the transaction signature. One pending claim per wallet; per-claim and daily caps apply. ## Things worth knowing * **Several tabs on one machine** compete for one GPU and do not earn more than one tab. The server's reassignment and plausibility checks make a fake second GPU pointless. * **A closed laptop lid** stops the tab. Set your machine not to sleep if you want it to keep working. * **Deploys reach you automatically.** A long-lived tab reloads itself at an idle moment when a new build ships. * **If the site shows LIVE DATA DELAYED**, the database is unreachable. Your tab keeps heartbeating and is not dropped; the figures catch up when it returns. * **The amount per hour is small and varies.** A fixed pool is split by everyone's verified work. More contributors means less each unless the operator raises the pool. The site shows the pool and the runway so you can see this for yourself. ## What earns nothing Being open without verifying units. Holding tokens without working. Reporting a better GPU than you have (it is ignored). Running the benchmark again (benchmarks are not paid). Browser inference stages that ran without a replica (recorded as `no-replica`). --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://docs.brainnetwork.app/participate/run-a-browser-node.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `build a script that syncs our docs to a CMS` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.