brain
No real nodes online right now REAL

One request.
The best available
intelligence.

BRAIN routes every request to the cheapest path that can run it, browser compute, cloud GPUs or external models, and attaches a receipt. Your computer can power it.

Use BRAIN→Power BRAIN
JOB #918282 assignedEXAMPLE
INSIDE NODE 8A21 · EXAMPLE JOB
GPUS ONLINE12,842■MEMORY68.4 TB■REQ/S482■INFERENCES 24H4.82M■QWEN 32B4,821 nodes■DEEPSEEK DISTILL2,184 nodes■EMBEDDINGS1,821 nodes■VISION884 nodes■CREATOR REWARDS 24H$18,482■PAID TO GPUS 24H$14,201■UPTIME99.91%■BRAIN$0.001842■SIM · DEMO DATA
GPUS ONLINE12,842■MEMORY68.4 TB■REQ/S482■INFERENCES 24H4.82M■QWEN 32B4,821 nodes■DEEPSEEK DISTILL2,184 nodes■EMBEDDINGS1,821 nodes■VISION884 nodes■CREATOR REWARDS 24H$18,482■PAID TO GPUS 24H$14,201■UPTIME99.91%■BRAIN$0.001842■SIM · DEMO DATA
SIM
Nodes online
—
Jobs completed
—
Verified work units
—
Compute units verified
—
Success rate
—
REAL · this server's records · refreshes every 15sOperations →

One request, traced.

A developer sends one request

POST /v1/chat/completions with model brain/auto. Same shape as the OpenAI API, so existing SDKs work unchanged.

01 / 03

One datacenter GPU.

The usual way to serve AI: a single 80 GB accelerator in a rack, rented by the hour.

GPUs
1
Memory
80 GB

1 point = 1 node · orange = serving a request now

Live topology

Four classes of supply, each shown in its real state. The graph below is the browser network: requests enter, split across device classes, execute, and merge. Green markers are real browsers connected right now; grey cells are simulated scale.

Browser compute
…
Verified parallel matmul on contributors' GPUs
Native GPU
…
Community machines running a native worker
Cloud GPU
…
Operator-controlled inference
External model
…
Third-party model API
Live topology
482 req/s0 liveSIM
    GPUs onlineSIM
    12,842
    Available memorySIM
    68.4 TB
    Inferences todaySIM
    4.82M
    Creator rewards todaySIM
    $18,482
    Paid to compute providersSIM
    $14,201
    Avg cost / 1M tokensSIM
    $0.XX
    Network uptimeSIM
    99.91%
    Power BRAIN

    Your computer
    can power BRAIN.

    No install, no driver, no CLI. Open a tab, let it measure your GPU, and join. The server issues a challenge only a real GPU can answer in time, then starts sending verifiable work.

    1. 01DetectReads exactly what WebGPU exposes. Anything it can’t see is labeled unavailable, never guessed.
    2. 02BenchmarkA WGSL kernel on your GPU. Score comes from the server’s clock, not yours.
    3. 03JoinYour node appears in the topology and starts receiving jobs within seconds.
    Power BRAIN→See live jobs
    BRAIN · NODE
    STATUS01/06
    DEVICE DETECTED
    WebGPU · READY
    adapter: apple · metal-3
    preview of the join sequence · values illustrative
    Rewards

    What your
    GPU earns.

    Rewards are paid for verified compute, from creator fees and inference sales. Holding tokens raises your multiplier, up to a cap. Holding alone earns nothing.

    Measure your GPU→How rewards work
    Your GPU19,500 score SIM
    Test your actual GPU

    About one second. Runs a real WGSL kernel; the server times and verifies the result.

    Presets are network median scores SIM · only the measured score is real

    Hours online per day20 h
    Token holdings2,500,000 BRAIN · 0.25%
    Estimated rewardEST
    $2.46/ day
    7 days $1730 days $74
    Where it comes fromcompute $1.48 · token boost +$0.98
    Token multiplier1.66× of 3× max
    1× no tokens3× cap
    Same GPU, no tokens
    $1.48 / day
    Same tokens, no GPU
    $0.00 · rewards require verified compute
    Contributor pool today
    $14,201 SIM

    Estimate using the production reward formula against a simulated network. Actual rewards depend on realized creator fees and inference sales, and on the verified compute of all other contributors. Not a guarantee of earnings.

    Use BRAIN

    One chat.
    Every receipt.

    Ask in the app or call the OpenAI-compatible API. BRAIN AUTO estimates every execution target, picks one for your mode and privacy, and tells you exactly where the answer ran and what it cost.

    • brain/autoAutomatically chooses the cheapest execution path that satisfies the request.
    • brain/qwenDistributed Qwen-class inference across the browser pool.
    • brain/codeCoding-optimized inference on distilled reasoning models.
    • brain/embedDistributed embeddings. Small model, highly parallel, browser-native.
    Open BRAIN chat→API reference
    POST /v1/chat/completions
    curl https://brainnetwork.app/v1/chat/completions \
  -H "Authorization: Bearer $BRAIN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "brain/auto",
    "messages": [{"role": "user", "content": "Explain this contract"}]
  }'

    Drop-in for OpenAI SDKs: set base_url to your BRAIN endpoint. Examples →