kenn-io/agentsview
Local-first session search, analytics, insights, and token use statistics for coding agents, supporting Claude Code, Codex, and more than 20 other agents.
What it solves
Agentsview provides a centralized way to browse, search, and track the costs of multiple AI coding agents. It eliminates the need to jump between different agent logs or interfaces to review past conversations, monitor token usage, or analyze spending across different tools.
How it works
The tool discovers session logs from various supported AI agents on your local machine, syncs them into a local SQLite database, and serves a web-based user interface. It can run as a background daemon to keep data updated via file watchers and provides a CLI for quick reports. For advanced setups, it supports PostgreSQL, DuckDB, and remote access via Quack.
Who it’s for
Developers who use multiple AI coding agents (such as Claude Code, Aider, Cursor, and others) and want a single pane of glass for their session history, cost tracking, and activity analytics.
Highlights
- Multi-Agent Support: Automatically discovers sessions from a wide variety of coding agents including Claude Code, Aider, Cursor, and Copilot CLI.
- Cost Tracking: Calculates token consumption and compute costs using LiteLLM and OpenRouter rates, including prompt-caching awareness.
- Search Capabilities: Offers full-text search (FTS5) and optional semantic search via OpenAI-compatible embeddings.
- Analytics Dashboard: Provides activity heatmaps, tool usage metrics, and session archetypes (e.g., automation vs. marathon sessions).
- Local-First: Session archives stay on the local machine by default.
- Developer Experience: Keyboard-first navigation and the ability to export sessions as HTML or GitHub Gists.
Related
- Project
- Project
- Project
- Project