termux-mcp-shell
Gives an agent full shell access on Termux with file read/write/edit capabilities.
README
termux-mcp-shell
MCP server (Streamable HTTP) that gives an agent full shell access on Termux, plus pi-grade file read/write/edit tools.
Install
pkg install python
pip install -r requirements.txt
Dependencies (see requirements.txt): only the mcp SDK is required. It pulls
in uvicorn + starlette (the ASGI runner for the Streamable HTTP transport),
plus httpx, anyio, pydantic, etc. as transitive deps. Everything else the
server uses is from the Python standard library.
Run
python server.py
Binds 0.0.0.0:8088 by default. MCP endpoint: /mcp.
- Local:
http://127.0.0.1:8088/mcp - LAN/hotspot:
http://<device-ip>:8088/mcp
Find the device IP with ifconfig or ip addr.
Configuration (env)
| Env | Default | Description |
|---|---|---|
MCP_HOST |
0.0.0.0 |
bind address |
MCP_PORT |
8088 |
port |
MCP_TRUNC_LIMIT |
8192 |
max stdout/stderr bytes per reply |
MCP_MAX_SESSIONS |
50 |
output buffers kept in memory (FIFO) |
MCP_READ_MAX_LINES |
2000 |
max lines per read_file |
MCP_READ_MAX_BYTES |
51200 |
max bytes per read_file (50KB) |
Tools
run_command(command, timeout?, cwd?): run via/bin/sh -c. Output larger thanMCP_TRUNC_LIMITbytes is truncated; the full output is kept in a buffer, get thesession_idfrom the result to read the rest.read_output(session_id, stream?, offset?, length?): read a continued chunk from a command's output buffer (stream:stdout/stderr).read_file(path, offset?, limit?): read a text file with 1-indexed line numbers.offset= start line,limit= number of lines. Auto-truncates at 2000 lines / 50KB and returnsnext_offsetto continue.write_file(path, content): create or overwrite a file; parent dirs are created automatically.edit_file(path, edits): exact-text replacement, modeled after pi.editsis a list of{old_text, new_text}. Eachold_textmust match exactly and be unique, is matched against the original content, must not overlap, and all edits apply atomically (all-or-nothing). If an exact match fails, a fuzzy match is tried (trailing whitespace, smart quotes, unicode dashes/spaces). CRLF line endings and BOM are preserved. Returns a unified diff.
Security note
No auth. Anyone on the same network can execute commands. Use only on a trusted network (your own device / private hotspot).
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.