Pyckle MCP connects your codebase to any MCP-compatible editor — Cursor, Windsurf, Continue, VS Code, and more. One subscription, every tool.
Switching from another AI coding tool? See how Pyckle compares →
The Model Context Protocol (MCP) is an open standard for connecting AI tools to external context sources. Pyckle implements it fully — meaning the same semantic search, session memory, and context routing that powers Claude Code is now available in every MCP-compatible editor.
No per-editor integrations. No duplicate config. One Pyckle subscription, and every AI tool you use gets the same deep codebase understanding.
Editor Compatibility
Pick your editor. Add your API key. Done.
// ~/.claude/claude_desktop_config.json { "mcpServers": { "pyckle": { "command": "pyckle", "args": ["mcp", "serve"], "env": { "PYCKLE_API_KEY": "YOUR_API_KEY" } } } }
For Cursor, Windsurf, and other IDE-based editors, use the SSE transport at
https://api.pyckle.co/mcp
— paste it into your editor's MCP server URL field.
17
tools available
3
transports (stdio, SSE, HTTP)
<50ms
context retrieval
Every tool available to Claude Code is available to any MCP client — the same codebase intelligence, everywhere.
search_code
Semantic code search by meaning, not keywords.
Learn more →index_codebase
Index your codebase for instant semantic queries.
Learn more →index_stats
View indexing statistics and health metrics.
Learn more →token_stats
Track token usage and optimize context budget.
Learn more →session_continue
Resume sessions with full context memory.
Learn more →session_summary
Generate summaries of current session activity.
Learn more →register_edit
Track file edits to keep context routing accurate.
Learn more →graph_neighbors
Explore import and dependency edges for any file.
Learn more →graph_impact
Analyze blast radius of code changes across the graph.
Learn more →index_obsidian
Index an Obsidian vault for cross-tool knowledge retrieval.
Learn more →index_notion
Index a Notion database or page for semantic queries.
Learn more →index_git_history
Query commit history with natural language.
Learn more →auto_context
Model-agnostic prompt routing for any MCP client.
Learn more →autoloop_init
Initialize an autonomous goal-directed iteration session.
Learn more →autoloop_log
Log iteration results — metric, commit, keep or discard.
Learn more →autoloop_status
Get progress summary and metrics for an active loop.
Learn more →autoloop_history
List all autoloop sessions and outcomes for a codebase.
Learn more →autoloop_complete
Mark an autoloop session as completed and archive results.
Learn more →add_memory
Save a decision, pattern, or insight to long-term memory.
Learn more →search_memory
Recall past decisions and patterns from memory by query.
Learn more →get_coverage
Fetch test coverage data for a file or the full codebase.
Learn more →index_git_issues
Index GitHub/GitLab issues for semantic issue search.
Learn more →Every prompt goes through three steps — invisible, automatic, <50ms.
Your codebase is chunked and embedded by PyckLM. The index lives in Pyckle's cloud, keyed to your API key — accessible from any editor.
Every prompt is intercepted by the MCP layer. Pyckle retrieves the most relevant code chunks using hybrid semantic + keyword search in ~50ms.
The right context is injected into the request before it reaches the model. Your AI gets exactly what it needs — nothing more, nothing less.
No add-on. No extra config. Subscribe once, connect everywhere.
Full MCP support ships with every Pyckle Pro subscription at $20/month. No separate purchase, no tier unlock — it's all just Pro.
$20/month or $200/year. Cancel anytime.