Show HN: Product analytics (and evals) for agent sessions on your MCP

4 hours ago

Armature captures and displays user sessions from AI agents, revealing user questions, agent reasoning, and usage patterns.
Sessions are grouped into use cases, ranked by volume and success rate, including unsupported use cases.
Models identify failures, loops, and dead ends in sessions, grouping them by root cause and ranking by user impact, even when APIs return 200 OK.
Each session is scored on goal achievement, allowing full replay of user asks, agent thinking, and all calls to pinpoint issues.
Integration is quick: sign up, generate an API key, paste a prompt for SDK wiring, and watch sessions appear within minutes.
Sessions are treated as sensitive data, with PII and secrets redacted by default before storage; custom needs can be addressed.

CueBench is a platform for evaluating and improving AI fluency among developers.
It analyzes sessions with AI coding assistants like Claude Code, Cursor, and Codex, providing scores and insights.
Key metrics include overall scores, risk signals (e.g., session loops), cost & efficiency, and model usage.

Coasty API enables autonomous task execution on managed machines, allowing users to run tasks, compose workflows, and access low-level prediction primitives for direct control when needed.
Authentication requires an API key sent via X-API-Key or Bearer headers, with test keys available for non-billing development and CI; keys must be kept secure and server-side.
Task runs involve creating runs with machine_id and task parameters, polling or using webhooks/SSE streams for status updates, and handling human takeovers via resume endpoints when agents pause.

Gemini Distillation Service trains a smaller 'student' model using outputs and reasoning from a larger 'teacher' model.
It bridges the gap between frontier model capabilities and production efficiency, offering lower latency and cost.
Unlike supervised fine-tuning (SFT), distillation leverages both final responses and internal reasoning paths from the teacher.

Telemetry.dev is an observability platform for AI and LLM apps, capturing every model call, tool step, and retrieval as OpenTelemetry traces with attached tokens, cost, latency, and errors.
Built on OpenTelemetry, it supports any framework and language without lock-in; first trace can be set up in five minutes.
Free plan offers 10,000 spans per month with no credit card required, 7-day retention, one project, and two seats.

<antirez>7 days ago

Redis Streams fill the gap between lists, sorted sets, and Pub/Sub for modeling message sequences or events.
The concept was inspired by Kafka consumer groups and a user's idea of a transaction log style data type.
Implementation uses radix trees and listpacks for space efficiency and indexed access by ID or time.

CueBench is a platform for evaluating and improving AI fluency among developers.
It analyzes sessions with AI coding assistants like Claude Code, Cursor, and Codex, providing scores and insights.
Key metrics include overall scores, risk signals (e.g., session loops), cost & efficiency, and model usage.

Coasty API enables autonomous task execution on managed machines, allowing users to run tasks, compose workflows, and access low-level prediction primitives for direct control when needed.
Authentication requires an API key sent via X-API-Key or Bearer headers, with test keys available for non-billing development and CI; keys must be kept secure and server-side.
Task runs involve creating runs with machine_id and task parameters, polling or using webhooks/SSE streams for status updates, and handling human takeovers via resume endpoints when agents pause.

Gemini Distillation Service trains a smaller 'student' model using outputs and reasoning from a larger 'teacher' model.
It bridges the gap between frontier model capabilities and production efficiency, offering lower latency and cost.
Unlike supervised fine-tuning (SFT), distillation leverages both final responses and internal reasoning paths from the teacher.

Telemetry.dev is an observability platform for AI and LLM apps, capturing every model call, tool step, and retrieval as OpenTelemetry traces with attached tokens, cost, latency, and errors.
Built on OpenTelemetry, it supports any framework and language without lock-in; first trace can be set up in five minutes.
Free plan offers 10,000 spans per month with no credit card required, 7-day retention, one project, and two seats.

<antirez>7 days ago

Redis Streams fill the gap between lists, sorted sets, and Pub/Sub for modeling message sequences or events.
The concept was inspired by Kafka consumer groups and a user's idea of a transaction log style data type.
Implementation uses radix trees and listpacks for space efficiency and indexed access by ID or time.

Hasty Briefsbeta