Bittensor service marketplace for agents
Discover and use Bittensor services.
One discovery layer for Bittensor. Find APIs, connect your agent and access services directly with x402, Gamma or your own API credentials.
Most integration-ready
Ranked by checks, documented access and documentation. No usage data.
- 1blockmachineSN1913
Live execution tested · Direct call documented · 1 operation with known parameters
- 2Decision ModelsSN9210
Direct call documented · 7 operations with known parameters · 2 supported access methods
- 3Green ComputeSN11010
Direct call documented · 19 operations with known parameters · 2 supported access methods
- 4LiumSN5110
Direct call documented · 19 operations with known parameters · 2 supported access methods
- 5AttelierSN1269
Direct call documented · 14 operations with known parameters · 1 supported access method
- 6CathedralSN949
Direct call documented · 38 operations with known parameters · 1 supported access method
- 7ChutesSN649
Direct call documented · 22 operations with known parameters · 1 supported access method
- 8EphemerisSN919
Direct call documented · 2 operations with known parameters · 1 supported access method
- Indexed subnets
- 129
- active on finney
- Available products
- 87
- researched from subnet sites
- API endpoints
- 2,990
- from validated OpenAPI specs
- Verified services
- 18
- API answers + valid spec
- MCP services
- 14
- publish an MCP server
Chain sync 19 Nov 2026, 14:54 UTC · block 9,541,345 · 8 on-chain service records. Counts come from the live registry; no usage or payment volumes are tracked.
Connect
Connect your AI agent
Hosted MCP · one URL, nothing to install
Read-onlyConnection commands are available at the public registry URL.
- Service search in plain words
- API docs and endpoints
- Access manifests
- Sign-up and payment methods
- Does not call providers
- Never takes keys, wallets or payments
Your agent then calls the provider directly with its own tools and your own access (API key, gamma deposit or x402 wallet). How it works
Explore
Service explorer
- ParetonReachableSN10 · pareton.ai
Searches LLM serving configurations (vLLM, TensorRT-LLM, SGLang) for cheaper engines that hold a latency cap
No access method documentedEndpoint reachable - FugalReachableSN70 · fugalai.com
LLM router that picks one model per question by price and quality, available as chat and an API in OpenAI and Anthropic formats
FreeAPI KeyEndpoint reachable - SayGMReachableSN28 · saygm.com
LLM gateway for Claude, GPT, Gemini and open models with OpenAI, Anthropic and Gemini API formats and private TEE inference
API KeyEndpoint reachable - Terza InferenceReachableSN103 · terza.si
Model API routing requests to miner-served models, with OpenAI- and Anthropic-style endpoints and billing on verified usage
API KeyEndpoint reachable - InstantNo APISN46 · instantnetwork.ai
Decentralized network of providers serving fast, uncensored LLM inference, with public provider performance stats
No access method documentedDocumented - ChutesVerifiedSN64 · chutes.ai
Serverless inference for open-source AI models, plus deployment of custom models on rented GPUs
API KeyAPI verified - Decision ModelsVerifiedSN92 · decisionmodels.io
Hosted API for open-weights decision models that return probabilities for each answer option to a question about a given state
API KeyAPI verified - ActualReachableSN95 · actual.inc
Runs open-source models on your own Mac, Windows or Linux machines behind one private API reachable through a relay
FreeAPI KeyEndpoint reachable - engyReachableSN53 · engy.ai
Inference API for open-source models served on a GPU network, compatible with OpenAI and Anthropic clients
API KeyGammaEndpoint reachable - YugosReachableSN34 · yugos.ai
Model routing API that picks a model per request across cost, speed and quality, with a CLI for coding agents
API KeyEndpoint reachable - Verathos DecideReachableSN96 · verathos.ai
Decision-model API returning probabilities for typed questions about a state, with signed receipts from the answering miner
API KeyEndpoint reachable - CacheonNo APISN14 · cacheon.ai
Hosted open-model inference and GPU serving-stack optimization built on an SGLang-based engine
No access method documentedDocumented - KubeTEENo APISN90 · kubetee.ai
Confidential AI inference and workloads on decentralized GPU clusters running inside trusted execution environments
No access method documentedDocumented - AttelierVerifiedSN126 · attelier.ai
Generative video, image, voice and music app with a pay-per-use inference API
API KeyAPI verified - ZilsReachableSN127 · zils.ai
Trains and hosts small decision models from labelled examples as a cheaper alternative to LLM calls
FreeAPI KeyEndpoint reachable - ChronoLLMNot checkedSN38 · chronollm.com
Point-in-time LLM vintages trained only on data up to each cutoff, for backtesting financial research without lookahead bias
FreeAPI KeyDocumented - IngotNo APISN97 · theingot.ai
Ternary-quantized GGUF build of a 27B open model, refined by subnet miners and distributed as a download for local inference
No access method documentedDocumented