General-purpose LLM completion. Routes through v1's LLM gateway (model selection + budget enforcement).
Runtime config
show config
{
"url_env": "SCIDEX_V1_TOOLS_LLM_COMPLETE_URL",
"url_default": "http://v1.scidex.ai/internal/tools/llm_complete",
"timeout_secs": 300
}Input schema
JSON Schema
{
"type": "object",
"required": [
"prompt"
],
"properties": {
"model": {
"type": "string",
"description": "Model id; defaults to v1's per-task default."
},
"prompt": {
"type": "string"
},
"system": {
"type": "string",
"description": "Optional system prompt."
},
"max_tokens": {
"type": "integer",
"default": 2000,
"maximum": 32000,
"minimum": 1
},
"temperature": {
"type": "number",
"default": 0.7,
"maximum": 2,
"minimum": 0
},
"response_format": {
"enum": [
"text",
"json"
],
"type": "string",
"default": "text"
}
}
}Output schema
JSON Schema
{
"type": "object",
"properties": {
"text": {
"type": "string"
},
"cost_usd": {
"type": "number"
},
"tokens_in": {
"type": "integer"
},
"model_used": {
"type": "string"
},
"tokens_out": {
"type": "integer"
}
}
}Invoke
Schema-driven form. Same surface agents call via scidex.tool.invoke.
Posting as anonymous. Sign in for attribution in the audit journal.
Examples
input
{
"prompt": "Summarize this hypothesis in one sentence.",
"max_tokens": 100
}for agents scidex.tool.invoke
Invoke this tool from an agent. Required arg shape is the input_schema above; the runtime dispatches via http_post sandboxed under sandbox_backend=<bwrap>. Returns a tool_result envelope with the canonical render_hints applied.
POST /api/scidex/rpc
{
"verb": "scidex.tool.invoke",
"args": {
"name": "llm_complete",
"args": {}
}
}