SuperAPI AI Directory
Partner with us中文AI newsFavorites
All toolsAI developer platformsdevrecall

devrecall

Local-first index of your commits, PRs, Jira/Linear tickets, Confluence docs, Slack threads and calendar events to search and cite past work.

Open website

About this tool

Purpose

DevRecall is designed to help developers aggregate and retrieve their own development activity records. It collects data from multiple commonly used tools and builds a local index, enabling you to conveniently review past work, generate reports, or query specific work content through a chat interface.

Main Capabilities

  • Pull data from Git, Slack, Google Calendar, Jira, Linear, Confluence, GitHub/GitLab/Bitbucket, Notion, and Claude Code sessions.
  • Store data in a local SQLite database, with full-text search support (requires the fts5 build tag).
  • Generate standup reports, weekly reports, and brag docs.
  • Provide chat functionality that answers questions based on your historical work records, with inline citations of commits, PRs, and tickets.
  • Include an MCP server that allows MCP-compatible programming tools (Claude Code, Cursor, Codex, Continue, Zed) to access your work memory through a stdio subprocess.
  • Support offline embeddings (ONNX + all-MiniLM-L6-v2), and can optionally use local Ollama or bring your own OpenAI/Anthropic key for chat.

Usage

Installation

  • macOS: Install the desktop app and CLI via Homebrew; the installation command is brew install pavelpilyak/tap/devrecall(not directly provided in the material, but inferred from the official text). The official text states "Install DevRecall.app and puts the devrecall CLI on your PATH.", but does not provide the specific command.
  • Linux / WSL: CLI and MCP server (no desktop app). The official text mentions "Without Homebrew, grab the tarball directly", supporting Ubuntu 22.04+ and equivalent versions (Debian 12+, Fedora 35+). The specific command is not provided.
  • Windows: No native build; all functionality except the desktop app can run under WSL, including the MCP server.
  • Build from source: Requires Go 1.22+ with CGO enabled (for SQLite FTS5). Build tags: fts5 enables full-text search; GO enables hugot's pure Go ONNX backend for embeddings.

Connect Data Sources and Generate Reports

The official text mentions "Then connect a source and generate your first standup", but does not give specific steps. See docs.devrecall.dev/install for the complete installation and setup guide.

MCP Server

Any MCP-compatible programming tool can launch DevRecall's MCP server as a stdio subprocess. For example, in Claude Code you can use /devrecall:recall what auth bug did I fix in February to query, returning cited commits, PRs, and tickets.

Use Cases

  • Developers need to quickly review their work history to prepare for standups or weekly reports.
  • Want to aggregate data from multiple development tools locally without relying on cloud services.
  • When using MCP-compatible tools, want to provide AI programming assistants with memory about delivered work.

Limitations

  • Data is stored entirely locally, with no cloud sync, so cross-device access must be handled on your own.
  • Windows has no native build and must run through WSL.
  • The desktop app is limited to macOS.
  • Building from source requires Go 1.22+ and CGO.
  • Some installation commands and specific configuration details are not provided in the official text; refer to docs.devrecall.dev.
  • Security: Tokens are stored in ~/.devrecall/tokens/(permissions 0600); the Cloudflare Worker relay is used only for OAuth callbacks and does not touch data.

Related tools

Browse category