Back to Apps

Distill
by Siddhant-K-code
Context intelligence layer for LLM agents providing persistent, deduplicated memory and context optimization.
0 stars
Works in:Cursor
Exposes:Tools
What it does
Distill acts as a deterministic intelligence layer that manages what an AI agent remembers across sessions. It eliminates redundant information via semantic deduplication, compresses aging context, and detects contradictory information to ensure the AI operates on a high-signal, trustable memory store.
Tools
store_memory: Save a piece of information with write-time deduplication and sensitivity tagging.recall_memory: Retrieve memories based on relevance and recency with task-relevance boosting.forget_memory: Remove outdated or deprecated memories.memory_expire: Soft-delete a memory without full removal.memory_supersede: Replace an old memory with a newer, updated version.memory_stats: Get statistics on the current memory store.create_session: Manage token-budgeted context windows for long-running tasks.push_session: Incrementally add content to a session with automatic budget enforcement.
Installation
Add the following to your claude_desktop_config.json:
{
"mcpServers": {
"distill": {
"command": "/path/to/distill",
"args": ["mcp", "--memory", "--session"],
"env": {
"OPENAI_API_KEY": "your-key"
}
}
}
}
Supported hosts
Confirmed for Claude Desktop and Cursor.
Quick install
go install github.com/Siddhant-K-code/distill@latestInformation
- Pricing
- free
- Published
- 6/12/2026
- stars
- 0
Categories
Choose your AI client and follow the steps below.
Cursor
Add mcp.json entry in ~/.cursor/ or project .cursor/Claude Desktop
{"mcpServers": {"distill": {"command": "/path/to/distill", "args": ["mcp", "--memory", "--session"], "env": {"OPENAI_API_KEY": "your-key"}}}}





