Skillget

docsagent

mcp
Security Audit
Warn
Health Warn
  • No license — Repository has no license file
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Community trust — 56 GitHub stars
Code Warn
  • network request — Outbound network request in src/index.ts
Permissions Pass
  • Permissions — No dangerous permissions requested
Purpose
This tool acts as a local-first document intelligence engine and MCP server. It allows AI agents and CLIs to securely index, search, and analyze private desktop files locally.

Security Assessment
The application inherently accesses sensitive data by reading your local documents to index them. However, the overall risk is Low. It does not request dangerous system permissions or execute arbitrary shell commands. While the README promises zero data leakage and absolute privacy, the automated scan flagged an outbound network request in the core source code. This request could simply be checking for updates, but it contradicts the strict "100% offline" marketing claims. Developers should quickly inspect the code to ensure no local file contents are ever transmitted. There are no hardcoded secrets present.

Quality Assessment
The project is actively maintained, with its most recent push occurring today. It has garnered a moderate level of community trust with 56 GitHub stars, indicating early adoption and positive peer validation. However, there is a discrepancy regarding its open-source status. The README displays an Apache-2.0 badge, but the repository lacks an actual license file, which leaves its legal usage rights undefined.

Verdict
Use with caution—verify the outbound network request in the source code before indexing highly sensitive files.
SUMMARY

DocsAgent is a local-first AI documents assistant that lets you securely index and chat with thousands of desktop documents with zero cloud leakage.

README.md

DocsAgent 🛡️

The High-Performance Local Intelligence Layer for AI Agents.

DocsAgent is a professional, local-first documents intelligence engine and MCP (Model Context Protocol) server. It provides a secure, near-instant bridge between your private local files and advanced agentic platforms like OpenClaw, Claude Code, and Cursor.

100% Local. 100% Private. Zero Data Leakage.

License: Apache-2.0
OS: macOS / Windows / Linux

🔥 Why DocsAgent?

Modern AI Agents are incredibly smart, but they are "blind" to your private desktop documents. DocsAgent acts as a High-Performance Intelligence Engine that indexes your local data, allowing agents to search, read, and analyze thousands of files (1,000+) with absolute privacy.

  • 🔒 Absolute Privacy: All document parsing, indexing, and vector storage happen strictly on your own hardware. Your data never leaves your machine.
  • ⚡ High-Performance Core: Powered by a native C++ engine. Experience millisecond-level retrieval even across massive document libraries.
  • 🔌 MCP Native: Built-in support for the Model Context Protocol (MCP), making it instantly compatible with the latest AI tools.
  • 📂 Format Mastery: Native parsing for PDF, Word (.docx), Excel (.xlsx), PPTX, Markdown, and TXT.
  • 🤖 Local LLM Ready: Works seamlessly with Claude 3.5/GPT-4o or run completely air-gapped with local models via Ollama.

🚀 Quick Start

1. Install via NPM

npm install -g @docsagent/docsagent

2. CLI Usage (Local Indexing & Search)

DocsAgent provides a powerful CLI for managing your local knowledge base.

  • Index your folders:
    # Add one or more directories or files to the index
docsagent add ~/Documents/Legal_Archive ~/Documents/Research
  • Search your documents:
    # Perform a semantic search directly from the terminal
docsagent search "What are the key terms in the Barclays contract?"

3. MCP Usage (Connect to AI Agents)

DocsAgent is a native Model Context Protocol (MCP) server. This allows AI agents to "see" your local documents.

Start the MCP server:

docsagent server

🔌 Agent Integration Examples

Connect DocsAgent to your favorite tools using the following configurations:

OpenClaw (Recommended)

OpenClaw is the most powerful UI for local-first agents. Add DocsAgent in your config.yaml:

mcpServers:
  docsagent:
    command: "docsagent"
    args: ["server"]

Claude Code

Integrate with Anthropic's CLI tool:

claude code mcp add docsagent -- docsagent server

Cursor / Windsurf

  1. Open Settings > Features > MCP Servers.
  2. Click + Add New MCP Server.
  3. Name: DocsAgent
  4. Type: command
  5. Command: docsagent server

🛠️ Agent Tools (via MCP)

Once integrated, your AI agent gains professional-grade document capabilities through these tools:

  • search: Perform deep semantic search across your entire local library to find relevant snippets.
  • add_docs: Instantly index new local folders or files during a conversation.
  • (More tools coming soon: list_documents, remove_document, status)

💻 CLI Reference

DocsAgent provides a powerful CLI with convenient aliases (dag, da):

Command Alias Example Description
server da server Start the MCP server service
add dag add <path> Add directories or files to the index
search da search "query" Search for documents directly from CLI
status dag status Check engine and indexing status
stop da stop Stop the background engine service

📂 Support Matrix

Format Status Features
PDF ✅ Full Deep layout analysis & high-speed parsing
Word (.docx) ✅ Full Table extraction & heading hierarchy preserved
PPTX ✅ Full Extracts slides, shapes, and speaker notes

🤝 Community & Contributing

We are building the open standard for Local RAG. Join us in making AI respect personal data sovereignty!

  • GitHub Issues: Found a bug? Have a feature request? Open an issue.
  • Star the Repo: If DocsAgent improved your workflow, please give us a ⭐ to help others find the project!

⚖️ License

DocsAgent is open-source under the Apache-2.0 License.

Desktop Sovereignty is here. Empower your Agents with DocsAgent.

Reviews (0)

No results found