Voker is an agent analytics platform that classifies intents, detects corrections, and measures resolution rates across AI agent conversations.
Key Takeaways
Lightweight Python/TypeScript SDK installs in two lines; works with OpenAI, Anthropic, Gemini, LangChain, CrewAI, and Vercel AI SDK.
Core primitives are Intents (user goals), Corrections (friction signals), and Resolutions (success rate) – structured from raw conversation data.
Self-service dashboards let PMs and analysts query conversation timelines without engineering tickets; targets teams with 1k+ chat sessions/month.
Pricing: free tier at 2k events/month; $80/mo Starter (20k events); $400/mo Agent First (2M events) with auto-optimization beta; enterprise gets self-hosted and SSO.
Integrates alongside Langfuse, LangSmith, PostHog, Mixpanel, and Amplitude rather than replacing existing observability stacks.
Hacker News Comment Review
The single comment asks how Voker differs from Langfuse, which also provides detailed agentic tracing – a direct competitive overlap Voker has not publicly addressed in the thread.
Notable Comments
@akslp2080: asks for differentiation from Langfuse on agentic tracing, a gap Voker’s own site sidesteps by listing Langfuse as a complementary tool.
