ContextPulse/contextpulse
by Various
Always-on context for AI agents: screen capture, OCR, voice, MCP server. Local-first, AGPL-3.0.
MCP
ContextPulse/contextpulse
Added 1 June 2026
Overview
ContextPulse is a local-first tool that provides always-on context for AI agents via screen capture, OCR, and voice input. It runs as an MCP server and is written in Python under the AGPL-3.0 license.
Best for
Best for
Developers building privacy-focused local AI agents that require continuous desktop context
Use cases
- Feeding real-time screen content to an AI coding assistant
- Automating desktop workflows with OCR-triggered actions
- Enabling voice-commanded local AI agents with visual context
How to use
Install
pip install -e packages/core -e packages/screen -e packages/voice -e packages/touch -e packages/project Tools exposed
get_screenshotget_recentget_screen_textget_monitor_summaryget_buffer_statusget_activity_summarysearch_historyget_context_atget_clipboard_historysearch_clipboardget_agent_statsget_recent_transcriptionsget_voice_statsget_vocabularyget_recent_touch_eventsget_touch_statsget_correction_historyidentify_projectget_active_projectlist_projects
Tested with
Claude Desktop, Claude Code, Cursor, Windsurf, VS Code
Notes
ContextPulse is a local-first tool that provides always-on context for AI agents via screen capture, OCR, and voice input. It runs as an MCP server and is written in Python under the AGPL-3.0 license.
0 stars on GitHub. Last updated 2026-05-25. Licensed AGPL-3.0.
Use cases
- Feeding real-time screen content to an AI coding assistant
- Automating desktop workflows with OCR-triggered actions
- Enabling voice-commanded local AI agents with visual context
Pros
- Local-first operation keeps data on device for privacy
- Multi-modal input (screen, OCR, voice) provides rich context
- Open source under AGPL allows modification and audit
Cons
- AGPL-3.0 license may be restrictive for commercial use
- Zero stars on GitHub indicates early-stage reliability
- Python implementation may have latency for real-time screen capture
Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.
Pros
- Local-first operation keeps data on device for privacy
- Multi-modal input (screen, OCR, voice) provides rich context
- Open source under AGPL allows modification and audit
Cons
- AGPL-3.0 license may be restrictive for commercial use
- Zero stars on GitHub indicates early-stage reliability
- Python implementation may have latency for real-time screen capture
Open-source & AI alternatives
Swap-in tools that solve the same job. Weigh the trade-offs before you commit.
Get the free Developer’s Field Guide
A 27-page field guide to the AI coding workflow with Claude. Claude Code, MCP servers, the prompt patterns that work, and what to delegate. Free.
Enter your work email. We send it straight over, plus a few short notes worth knowing. Unsubscribe any time.
