Use Cases
What You Can Build
TensorGate gives AI agents direct access to Bittensor's decentralized intelligence network — plus a mesh network where agents discover and call each other’s tools. No API keys. No vendor lock-in. Just tool calls.
The Opportunity
Why Decentralized AI Matters
Centralized AI has single points of failure. Bittensor replaces them with market incentives.
No API Keys Required
Query 50+ specialized AI subnets without vendor accounts, rate limits, or billing dashboards. TensorGate connects to public chain endpoints.
Censorship Resistant
No single provider can shut down your agent's intelligence. Inference is distributed across thousands of miners globally.
Market-Driven Quality
Miners compete for TAO rewards. Validators stake on the best performers. Economic incentives drive better AI, not corporate roadmaps.
Real Examples
See It in Action
Concrete agent-to-TensorGate interactions. Each example shows the exact tool calls an AI agent makes.
01 — Market Intelligence Agent
Analyze tokens and get predictions
An agent combines token analysis with SN8 prediction data to form a market thesis. Two tool calls, zero configuration.
tgate_analyze_token
tgate_predictions
$ Agent: "Analyze TAO token and get price predictions"
→ tgate_analyze_token({ token_symbol: "TAO" })
Price: $325.42
24h: +4.2%
Market Cap: $2.4B
→ tgate_predictions({ asset: "TAO", timeframe: "7d" })
Subnet: SN8 (Time Series Prediction)
Miners: 4,096 competing
Method: Yuma consensus aggregation
→ tgate_analyze_token({ token_symbol: "TAO" })
Price: $325.42
24h: +4.2%
Market Cap: $2.4B
→ tgate_predictions({ asset: "TAO", timeframe: "7d" })
Subnet: SN8 (Time Series Prediction)
Miners: 4,096 competing
Method: Yuma consensus aggregation
02 — Network Monitor
Track Bittensor health in real time
A monitoring agent checks overall network status and drills into specific subnet performance. Useful for validators, stakers, and researchers.
tgate_network_stats
tgate_subnet_health
$ Agent: "Check Bittensor network status"
→ tgate_network_stats()
TAO Price: $325.42
Block Height: 3,847,291
Subnets: 52 active
Total Stake: 6.2M TAO
Emission: 7,200 TAO/day
→ tgate_subnet_health({ subnet_id: 1 })
Subnet: SN1 — Text Prompting
Miners: 256 active
Top Miner: 0.0847 incentive
Emission: 312 TAO/day
→ tgate_network_stats()
TAO Price: $325.42
Block Height: 3,847,291
Subnets: 52 active
Total Stake: 6.2M TAO
Emission: 7,200 TAO/day
→ tgate_subnet_health({ subnet_id: 1 })
Subnet: SN1 — Text Prompting
Miners: 256 active
Top Miner: 0.0847 incentive
Emission: 312 TAO/day
03 — Research Agent
Route research queries to specialized subnets
An agent sends a summarization task and TensorGate auto-routes to SN1 (Text Prompting), querying top miners by incentive score.
tgate_query_subnet
$ Agent: "Summarize recent advances in protein folding"
→ tgate_query_subnet({
prompt: "Summarize recent advances in protein folding",
task_type: "summarize"
})
Routed to: SN1 (Text Prompting)
Miners queried: 12 (top by incentive)
Consensus: Yuma-weighted aggregation
→ tgate_query_subnet({
prompt: "Summarize recent advances in protein folding",
task_type: "summarize"
})
Routed to: SN1 (Text Prompting)
Miners queried: 12 (top by incentive)
Consensus: Yuma-weighted aggregation
04 — Multi-Subnet Pipeline
Chain subnets for complex workflows
Generate an image on SN9, then analyze it on SN19. TensorGate makes cross-subnet pipelines trivial — each step is just another tool call.
tgate_query_subnet (SN9)
tgate_query_subnet (SN19)
$ Agent: "Generate and analyze a product mockup"
→ tgate_query_subnet({
prompt: "Minimalist SaaS dashboard, dark theme",
task_type: "text-to-image",
subnet_id: 9
})
Routed to: SN9 (Text to Image)
→ tgate_query_subnet({
prompt: "Describe the UI layout and suggest improvements",
task_type: "vision",
subnet_id: 19
})
Routed to: SN19 (Vision)
→ tgate_query_subnet({
prompt: "Minimalist SaaS dashboard, dark theme",
task_type: "text-to-image",
subnet_id: 9
})
Routed to: SN9 (Text to Image)
→ tgate_query_subnet({
prompt: "Describe the UI layout and suggest improvements",
task_type: "vision",
subnet_id: 19
})
Routed to: SN19 (Vision)
05 — Staking Advisor
Recommend delegation strategies
An agent pulls validator data and staking metrics to help users make informed delegation decisions.
tgate_staking_info
$ Agent: "What are the top validators for staking?"
→ tgate_staking_info()
Total Staked: 6.2M TAO
Active Validators: 1,024
Avg APY: ~14.2%
Min Stake: No minimum for delegation
→ tgate_staking_info()
Total Staked: 6.2M TAO
Active Validators: 1,024
Avg APY: ~14.2%
Min Stake: No minimum for delegation
06 — Miner Benchmark
Rank miners across subnets
Benchmark miner performance across multiple subnets at once. Identify top performers, track trends over time, and discover miners active on multiple subnets.
tgate_subnet_benchmark
$ Agent: "Benchmark SN1 and SN8 miners with trends"
→ tgate_subnet_benchmark({
subnets: [1, 8],
limit: 5,
include_history: true
})
SN1 Top Miner: UID 42 — 8.47% incentive (Rising)
SN8 Top Miner: UID 17 — 6.12% incentive (Stable)
Cross-Subnet: 3 miners active on both
→ tgate_subnet_benchmark({
subnets: [1, 8],
limit: 5,
include_history: true
})
SN1 Top Miner: UID 42 — 8.47% incentive (Rising)
SN8 Top Miner: UID 17 — 6.12% incentive (Stable)
Cross-Subnet: 3 miners active on both
07 — Tool Discovery
Find any tool across the MCP ecosystem
Search the decentralized tool registry by capability. TensorGate returns matching tools from all registered MCP servers, ranked by trust score and health status. The npm registry, but for AI tools.
tgate_discover
$ Agent: "Find tools that can send emails"
→ tgate_discover({ query: "send email", category: "communication" })
Found: 3 tools across 2 servers
1. send_email (my-email-server) — Trust: 85% • healthy
2. gmail_send (gmail-mcp) — Trust: 72% • healthy
3. smtp_relay (comms-hub) — Trust: 68% • degraded
→ tgate_discover({ query: "send email", category: "communication" })
Found: 3 tools across 2 servers
1. send_email (my-email-server) — Trust: 85% • healthy
2. gmail_send (gmail-mcp) — Trust: 72% • healthy
3. smtp_relay (comms-hub) — Trust: 68% • degraded
08 — Cross-Server Mesh Call
Call tools on remote MCP servers
After discovering a tool, call it directly through the mesh network. TensorGate proxies the request through the hub, tracks latency and success rate for trust scoring, and returns the result.
tgate_discover
tgate_mesh_call
$ Agent: "Send email to alice@acme.com about the TAO report"
→ tgate_mesh_call({
server_id: "abc-123",
tool_name: "send_email",
arguments: { to: "alice@acme.com", subject: "TAO Report", body: "..." }
})
Server: my-email-server (trust: 85%)
Latency: 450ms
Result: Email sent successfully
→ tgate_mesh_call({
server_id: "abc-123",
tool_name: "send_email",
arguments: { to: "alice@acme.com", subject: "TAO Report", body: "..." }
})
Server: my-email-server (trust: 85%)
Latency: 450ms
Result: Email sent successfully
09 — Natural Language Routing
One sentence, automatic tool chain
Describe what you want in plain English. TensorGate searches the registry, picks the best tool by trust + relevance, extracts arguments from your intent using heuristic pattern matching, and calls it. DNS for AI.
tgate_mesh_ask
$ Agent: "What is the current TAO price?"
→ tgate_mesh_ask({ intent: "what is the current TAO price" })
Discovered: tgate_network_stats (tensorgate)
Trust: 90% • healthy • verified
Latency: 320ms
TAO Price: $412.53 (+3.2%)
Market Cap: $2.8B
24h Volume: $89M
→ tgate_mesh_ask({ intent: "what is the current TAO price" })
Discovered: tgate_network_stats (tensorgate)
Trust: 90% • healthy • verified
Latency: 320ms
TAO Price: $412.53 (+3.2%)
Market Cap: $2.8B
24h Volume: $89M
10 — Deep Research
One call, full intelligence report
The research meta-tool auto-detects what data you need and chains multiple tools in parallel. Get a synthesized report with executive summary in a single tool call.
tgate_research
$ Agent: "Deep research on TAO"
→ tgate_research({ topic: "TAO", depth: "deep" })
Tools chained: 5 (stats, token, predict, stake, bench)
Price: $325.42 (+4.2%)
Total Stake: 6.2M TAO
SN8 Top Miner: UID 17 — 6.12% incentive
Report: Executive summary + 5 sections
→ tgate_research({ topic: "TAO", depth: "deep" })
Tools chained: 5 (stats, token, predict, stake, bench)
Price: $325.42 (+4.2%)
Total Stake: 6.2M TAO
SN8 Top Miner: UID 17 — 6.12% incentive
Report: Executive summary + 5 sections
Why MCP
The MCP Advantage
One protocol. Every subnet. No custom integration work.
Traditional Approach
Custom SDK Per Provider
- Separate SDK for each AI provider
- API keys and billing per service
- Rate limits and usage caps
- Single point of failure
- Provider can censor or shut down
TensorGate Approach
One MCP Config
- Single MCP server, 19 tools, 50+ subnets + mesh network
- Zero API keys required
- No rate limits — query the chain directly
- Distributed across thousands of miners
- Censorship resistant by design
Compatibility
Works With Your Agent
Any MCP-compatible client discovers TensorGate's tools automatically.
Claude Code
CLI agentClaude Desktop
Desktop appCursor
AI IDECustom Agents
Any MCP clientStart building.
Install TensorGate and query decentralized AI in under a minute.