Context Mode
Reduces context consumption by 98% through sandboxed code execution and SQLite-based session continuity with FTS5 search.
Context Mode is an MCP server providing sandboxed code execution in 11 languages and session continuity tracking. It keeps raw data out of the context window through isolated subprocess execution, reducing context usage from 315 KB to 5.4 KB in typical sessions. Session events are persisted in SQLite with FTS5 indexing, allowing the model to resume seamlessly after context compaction without losing file edits, tasks, or git operations.
Source
Repository: https://github.com/mksglu/context-mode
Maintain Context Mode?
Let people know it's listed here — add the badge (live metrics, light/dark aware) or a plain link to your README or docs.
[](https://getagentictools.com/mcp/mksglu-context-mode?ref=badge) Related mcp
Claude Context Provides semantic code search and indexing using vector embeddings and AST-based code splitting, enabling natural language queries
Assistant UI The MCP server tools have been designed to allow an agent to query the specific information it needs to complete an assistant-ui r
PraisonAI Multi-agent framework with self-reflection capabilities
Airweave Search Search Airweave collections with natural language queries.
Cognee AI-friendly database and knowledge-management capabilities via various database schemes.
Mastra Docs Provides AI assistants with direct access to Mastra.ai's complete knowledge base.