LLM completion

1 curl -X POST https://api.asi1.ai/v1/chat/completions \ 2 -H "Authorization: Bearer <api_key>" \ 3 -H "Content-Type: application/json" \ 4 -d '{ 5 "messages": [ 6 { 7 "content": "What are you capable of?", 8 "role": "user" 9 } 10 ], 11 "model": "asi1-mini" 12 }'

1 { 2 "choices": [ 3 { 4 "finish_reason": "stop", 5 "delta": { 6 "content": "I am capable of searching agents on agentverse. I only come when stream is true.", 7 "role": "user", 8 "tool_calls": [ 9 { 10 "id": "toolcall-abc123", 11 "tool_name": "web_search", 12 "args": {}, 13 "result": {} 14 } 15 ] 16 }, 17 "index": 0, 18 "message": { 19 "content": "I am capable of searching agents on agentverse. I only come when stream is false", 20 "role": "user", 21 "tool_calls": [ 22 { 23 "id": "toolcall-abc123", 24 "tool_name": "web_search", 25 "args": {}, 26 "result": {} 27 } 28 ] 29 } 30 } 31 ], 32 "id": "cmpl-7aNwD", 33 "usage": { 34 "completion_tokens": 200, 35 "prompt_tokens": 100, 36 "total_tokens": 300 37 }, 38 "executable_data": [ 39 {} 40 ], 41 "metadata": {} 42 }

Handles client LLM completion requests, applies rate limits based on tokens and requests, supports streaming or normal responses, and records usage.

Request

LLM Completion Payload

messageslist of objectsRequired

modelstringRequired

agent_addressstringOptional

max_tokensintegerOptional

planner_modebooleanOptional

streambooleanOptional

study_modebooleanOptional

temperaturedoubleOptional

toolslist of objectsOptional

web_searchbooleanOptional

Response

Successful completion or stream

choiceslist of objects

A list of generated choices/messages.

idstring

usageobject

Information about token usage.

executable_datalist of objects or null

metadataobject or null

Headers

Request

Response

Errors