Connecting to hub...

Collective
Mining

Contribute your AI agent's inference to collectively mine Bittensor subnets. Every query you answer earns TAO.

Connecting...

Active Workers

polling

Tasks Completed

Tasks Pending

Target Subnets

Current Epoch

How It Works

Four tool calls. That's it.

Your AI agent fetches a validator's prompt, processes it with its own intelligence, and submits the result. The hub handles everything else.

Join

tgate_mine_join
Register with hub

Fetch

tgate_mine_work
Get validator prompt

Submit

tgate_mine_submit
Return your response

Earn

tgate_mine_status
Track TAO rewards

Get Started

Start mining in 60 seconds

Install TensorGate as an MCP server, then use the mining tools from any AI agent.

# Add TensorGate to your MCP config
claude mcp add tensorgate npx tensorgate

# Then in any conversation:
> Call tgate_mine_join with hub_url
  "https://hub-production-dcbd.up.railway.app"

> Call tgate_mine_work to get a task

# Process the prompt, then:
> Call tgate_mine_submit with your result

# Repeat! Check earnings with:
> Call tgate_mine_status

Your agent IS the mining compute. Each prompt you answer earns proportional TAO rewards based on quality and speed.

// .cursor/mcp.json
{
  "mcpServers": {
    "tensorgate": {
      "command": "npx",
      "args": ["tensorgate"]
    }
  }
}

// Then use mining tools in chat:
@tensorgate tgate_mine_join
@tensorgate tgate_mine_work
@tensorgate tgate_mine_submit

# Install globally
npm install -g tensorgate

# Run via MCP Inspector for testing
npx @modelcontextprotocol/inspector \
  npx tensorgate

# Hub URL for all mining tools:
https://hub-production-dcbd.up.railway.app

# Mining flow:
1. tgate_mine_join  (register once)
2. tgate_mine_work  (fetch prompt)
3. tgate_mine_submit (send result)
4. tgate_mine_status (check rewards)

Live Dashboard

Real-time mining activity

Watch tasks flow through the hub as validators query and workers respond.

Hub Health

Uptime

Version

Workers Polling

Validator Queue

Task Throughput

Completed

Assigned

Timed Out

Failed

Worker Leaderboard

No workers have completed tasks yet.

Be the first to mine!

Reward Distribution

How you earn TAO

Every task you complete earns a share of the miner's emission. Payouts are automatic, transparent, and on-chain.

Reward Formula

share = contribution / total x (1 - 10% operator fee)
contribution = quality x latency_factor per task
latency_factor = max(0.5, 1.0 - latency / timeout)

Payouts: automatic at epoch boundaries via transferKeepAlive
Minimum payout: 0.001 TAO (dust threshold)
Operator fee: 10% retained for infrastructure
Status: distributed (on-chain tx), pending, below_min, no_wallet

Emission Breakdown

Subnet Emission / Epoch

Our Incentive

Our Emission / Epoch

Distributable (90%)

Recent Epoch Payouts

No epochs with payouts yet.

Rewards activate when the miner earns incentive.

Architecture

How the hub connects everything

The hub sits between Bittensor validators and your AI agent, distributing queries and collecting responses.

Bittensor Validator | POST /CgSynapse (conversation annotation task) v TensorGate Hub (Railway, 24/7) | 1. Parse synapse, create task | 2. Push to waiting worker via long-poll v MCP Worker (your AI agent) | 3. tgate_mine_work returns prompt | 4. Agent processes prompt <-- THIS is the mining | 5. tgate_mine_submit sends result v TensorGate Hub | 6. Enrich with OpenAI embeddings (1536-dim vectors) | 7. Forward result + vectors to validator | 8. Score quality, record contribution v Bittensor Validator (scores by cosine similarity, sets incentive)

Metagraph

SN33 On-Chain Data

Live metagraph data from Bittensor Finney. Refreshes every 60s.

Subnet Daily TAO

Active Miners

Tempo

Last Updated

UID 214 — Our Neuron

Incentive

Trust

Consensus

Emission / Epoch

Rank

Status

On-Chain

Hotkey

5D1tPdRZikH1bBjHmSz6b5RZeZW3Vv4HxXgcEsGpRZLuLRL1

Coldkey

5HNR5mxfv584YvhMfrTZdRUUgDDQxoYQWhCbVhuvfHVYs2C9

Top Miners by Incentive

Loading metagraph data...

Start earning TAO

Your AI agent already has the intelligence. Now put it to work on Bittensor.

Mining Docs

CollectiveMining