ByteZone
Back to listClaim this project

vshulczDeveloper tools

Deja Vu

deja-vu

A local memory layer for coding agents that indexes and retrieves past session decisions and code edits.

  • Backend
  • Full stack
  • DevOps
  • Automation
  • Productivity
  • Desktop
  • CLI
  • Windows
  • macOS
  • Linux
  • Self-hostable
  • Runs locally
  • Docker supported
Open websiteView on GitHub
deja-vu screenshot
Popularity
723 Stars
GitHub stars
Recent activity
8/26/2026
Updated in the last 30 days
License
MIT
Permissive

Why it matters

We look beyond stars: what problem it solves, whether it creates real utility, and what makes its approach worth noticing.

Last 90 days

Problem

AI coding agents typically lack long-term memory across sessions, forcing users to repeatedly solve the same problems and wasting time and compute.

Practical value

By indexing existing local session logs without requiring LLM calls or vector databases, it enables efficient context retrieval and significantly improves agent continuity.

Innovation / differentiation

It bypasses standard vector embedding approaches in favor of a lightweight local binary index, leveraging existing agent history files as a knowledge base with high accuracy and performance.

Leverage potential

Supports various mainstream coding agents via MCP and plugin integration, automatically injecting historical decisions into new sessions and reducing the burden of manual context management.

Why now

As AI coding agents become widespread, effectively utilizing historical interaction data has become a key pain point for productivity, making this tool highly relevant.

Community activity

In the last 90 days there were 813 new issues and 1160 pull requests; the bounded issue/PR samples include 11 issue authors and 5 PR contributors, with 49 releases.

Maintainer responsiveness

The 100-issue window sample had a 98% close rate, and the 100-pull-request sample had a 95% merge rate. Maintainer-response observations covered 100% of that issue sample, with a 18% response rate and median first response of 10.6 hours.

5 contributors in PR sample49 releasesIssue response rate 18% · sample 100 (100% coverage)Median first response 10.6hPR merge rate 95% · 100 sampled in window

Key highlights

  • Automatically indexes local session history
  • Provides memory recall to agents via MCP
  • Supports decision state tagging (accepted/rejected)

Quick start

How it is installed, how hard it is, and where to start.

Where it runs

Runs locally

Difficulty

Easy — follow the steps

Docker supportedSelf-hostableRuns locally
  1. 01Run the installation script to deploy the binary.
  2. 02Execute the install command to automatically configure agent integrations.
  3. 03Invoke memory features within your coding agent via MCP or plugins.

Best for

  • Teams that want data on their own servers
  • Developers who want to try it on their machine
  • People who prefer Docker deploys

More about it

deja-vu is a local binary that turns the vast history of coding agent sessions (e.g., Claude Code, Aider, Cursor) into a searchable memory layer. By indexing the raw session data written to disk, it automatically injects past decisions and solutions into new sessions or prompts.

It operates entirely locally without requiring cloud APIs or embedding models. Through the MCP protocol, it integrates with various coding agents, ensuring they work based on previous successful outcomes rather than re-debugging.

Sources

Each field shows its status and source — expand to review.

15 · Expand

  • capability tags

    Verified

    automation, productivity

    Source: admin_cms · cms editor · 8/22/2026

  • editor note

    Verified

    {"en":"This tool provides instant memory by reading local agent session logs without requiring training. It is ideal for developers who switch between coding agents or need to reuse past debugging decisions.","zh":"该工具通过直接读取代理在本地生成的会话日志，无需额外训练即可实现即时记忆。它特别适合需要频繁在不同编码代理间切换，或希望复用过往调试决策的开发者。"}

    Source: admin_cms · cms editor · 8/22/2026

  • how to use

    Verified

    {"steps":[{"en":"Run the installation script to deploy the binary.","zh":"运行安装脚本以部署二进制文件。"},{"en":"Execute the install command to automatically configure agent integrations.","zh":"执行安装命令以自动配置代理集成。"},{"en":"Invoke memory features within your coding agent via MCP or plugins.","zh":"在编码代理中通过 MCP 或插件调用记忆功能。"}],"installAt":"local","difficulty":"easy"}

    Source: admin_cms · cms editor · 8/22/2026

  • intro

    Verified

    {"en":"deja-vu is a local binary that turns the vast history of coding agent sessions (e.g., Claude Code, Aider, Cursor) into a searchable memory layer. By indexing the raw session data written to disk, it automatically injects past decisions and solutions into new sessions or prompts.\n\nIt operates entirely locally without requiring cloud APIs or embedding models. Through the MCP protocol, it integrates with various coding agents, ensuring they work based on previous successful outcomes rather than re-debugging.","zh":"deja-vu 是一个本地运行的二进制工具，将编码代理（如 Claude Code, Aider, Cursor 等）产生的海量历史会话转化为可检索的记忆层。它通过索引代理写入磁盘的原始会话数据，在新的会话开始或用户提问时，自动注入过往的决策与解决方案。\n\n该工具无需云端 API 或嵌入模型，所有数据处理均在本地完成。它支持通过 MCP 协议与多种编码代理集成，确保代理能够基于之前的成功经验进行工作，而非重复调试。"}

    Source: admin_cms · cms editor · 8/22/2026

  • Latest release

    Verified

    v0.18.0

    Source: GitHub API · latest_release=v0.18.0 · 8/26/2026

  • License

    Verified

    MIT

    Source: GitHub API · license.spdx_id=MIT · 8/26/2026

  • needs api key

    Verified

    No

    Source: admin_cms · cms editor · 8/22/2026

  • One-liner

    Verified

    {"en":"A local memory layer for coding agents that indexes and retrieves past session decisions and code edits.","zh":"一个本地运行的编码代理记忆层，可自动索引并检索历史会话中的决策与代码修改。"}

    Source: admin_cms · cms editor · 8/22/2026

  • Platforms

    Verified

    windows, macos, linux

    Source: admin_cms · cms editor · 8/22/2026

  • Category hint

    Inferred from materials

    ai-apps

    Source: Project README · hint=ai-apps · 8/15/2026

  • product forms

    Verified

    desktop, cli

    Source: admin_cms · cms editor · 8/22/2026

  • role tags

    Verified

    backend, fullstack, devops

    Source: admin_cms · cms editor · 8/22/2026

  • supports docker

    Verified

    Yes

    Source: Repository file · dockerfile=true; compose=false · 8/26/2026

  • supports local

    Verified

    Yes

    Source: admin_cms · cms editor · 8/22/2026

  • supports self host

    Verified

    Yes

    Source: admin_cms · cms editor · 8/22/2026

Related projects

Other verified projects matched by category, capabilities, and intended roles.

WSL

Windows Subsystem for Linux allows you to run Linux command-line tools and applications directly on Windows without the overhead of a traditional virtual machine.

Same Developer tools categoryBoth support Automation

opencodex

A universal local proxy that enables the use of any LLM, such as Claude, Gemini, or DeepSeek, within Codex and Claude Code, while managing ChatGPT account pools.

Same Developer tools categoryBoth support Automation

cli-printing-press

Automatically generates agent-optimized Go CLI tools and MCP servers by reading API documentation or sniffing web traffic.

Same Developer tools categoryBoth support Automation

agentdock

Secure MCP runtime for AI agents to operate local machines, servers, and containers with multi-device orchestration.

Same Developer tools categoryBoth support Automation

omni

Reduces AI Agent input token consumption by replacing redundant file reads and command outputs with compact, retrievable handles.

Same Developer tools categoryBoth support Automation

massing

Open-source, self-hosted, IFC-native AEC platform featuring in-browser BIM modeling, a general contractor portal, and development proforma tools.

Same Developer tools categoryBoth support Automation