A local voice assistant gateway providing real-time voice interaction, LLM integration, WebRTC support, and agent task execution with built-in tools for web search, math, time, and identity queries.
OpenLive Gateway exposes 4 tools with minimal schema documentation. Tool definitions are scattered across JavaScript and Rust files without centralized registration. Only web_search and calculator have visible input schemas (single string parameter each); get_time and identity lack explicit schema definitions in the source. Descriptions are present but generic (10-40 chars), well below the 50-200 char LLM-optimized baseline. No parameter descriptions, no output schemas, no error handling guidance. Tools lack verb-noun naming clarity (identity is ambiguous). No evidence of tool annotations, pagination, or composition patterns.
Math evaluation and simple arithmetic computation
Get current UTC time
Self-identification response for 'who are you' queries
Web search for fact-ish queries and information lookup
get_time and identity lack visible input schemas in source code. Cannot verify parameter types or constraints.
All tool descriptions are under 20 characters ('Get current UTC time', 'Web search', 'Math evaluation', 'Self-identification response'). Too short to guide LLM selection; lack context on when/why to use each tool.
No parameter descriptions visible. 'query' and 'expression' parameters lack guidance on format, constraints, or examples. 'intent' parameter in identity tool is undocumented.
Tool 'identity' uses ambiguous naming. Does not start with action verb (get_, search_, etc.). Unclear what 'identity' does vs. other tools.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 40 | 2026-07-28+ | v2 |
No output schemas documented. LLMs cannot infer what fields to expect from tool results, blocking downstream tool chaining and data extraction.