Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.agnost.ai/llms.txt

Use this file to discover all available pages before exploring further.

Agnost AI is an analytics platform for AI agents that understands what your users actually want: and how they feel about your agent. It reads every conversation, extracts intent signals and sentiment, surfaces frustration patterns, and uses those insights to make your agents better every day. Most observability tools tell you a span took 3.2 seconds. Agnost tells you the user was frustrated, asked the same thing three times, and dropped off: and which tool call caused it.

​
Signals: what users actually want

Every message is classified into intents and sentiment, then aggregated. The Signals dashboard shows you, in plain English, what users keep asking for, what they love, what they hate, and what’s missing from your product. Agnost AI Signals dashboard You’ll see:
  • Intent clusters: automatic grouping of every conversation by what the user was trying to do.
  • Sentiment over time: how user satisfaction trends day-by-day, broken down by intent or feature.
  • Hidden feature requests: things users keep asking for that you don’t yet ship.
  • Recurring frustrations: repeated questions, abandoned flows, failed tool calls, refusals, and dead ends.

​
How it fits in your stack

Your AI Agent  →  Agnost AI  →  Signals · Evals · Improvements
Agnost is OpenTelemetry-native and works with any LLM and any framework. Two-minute setup. See the sidebar for SDK and framework integrations.