Explorer
Syncing—

Text Generation - OpenAI compatible LLM inferenceService

text-generation

OpenAI-API-compatible text generation. The expected endpoints are POST /v1/chat/completions and GET /v1/models; suppliers MAY expose more of the OpenAI surface but consumers MUST NOT depend on it. Address the model as "pocket_network". Each supplier brings its own model and weights, so response quality varies and is not guaranteed by the protocol.

Active Suppliers
5
of 5 ever configured
Service Info
Service ID
text-generation
Compute Units / Relay
300,000
Owner
pokt1l…2h0l
Active Applications
3
Relay Mining EMA
37 relays
Service Metadata Cardsource: on-chain metadata (LCD)
{8 items
"schema":"pocket-service-card/v1"
"description":"OpenAI-API-compatible text generation. The expected endpoints are POST /v1/chat/completions and GET /v1/models; suppliers MAY expose more of the OpenAI surface ..."
"rpc_types":[1 item
0:{4 items
"type":"REST"
"intent":"expected"
"backend_hint":"OpenAI-compatible server on :8000 (vLLM); mount at /"
"notes":"Only POST /v1/chat/completions and GET /v1/models are expected. Token streaming is SSE over the same POST ("stream": true), not a separate transport."
}
]
"apis":[2 items
0:"openai-chat-completions"
1:"openai-models"
]
"specs":[1 item
0:{3 items
"kind":"markdown"
"url":"https://raw.githubusercontent.com/pnyxai/pocket-network-services/88b583cbfc60f44edc8adaee2bae59f62a4e5f94/mainnet/docs_text-generation.md"
"sha256":"d83f2b125a38d68b3274d996718bcdea42c78bea0939206d8993f86b35490f52"
}
]
"access":"public"
"results":"variable"
"serving":{3 items
"backend":"Any OpenAI-compatible inference server exposing POST /v1/chat/completions and GET /v1/models. Bring your own model; no specific weights are mandated. The served..."
"healthcheck":[1 item
0:{3 items
"rpc_type":"REST"
"request":{3 items
"path":"/v1/chat/completions"
"method":"POST"
"body":{4 items
"model":"pocket_network"
"messages":[1 item
0:{2 items
"role":"user"
"content":"Reply with the single word: pong"
}
]
"max_tokens":8
"temperature":0
}
}
"expect":{2 items
"json_path":"$.choices[0].message.content"
"matches":"(?i)pong"
}
}
]
"notes":"Use a server with continuous batching so one backend serves many concurrent generations. Alias the served model as "pocket_network". It is recommended to isolat..."
}
}