#local-first

Strata is a self hosted AI memory server. Your AI remembers everything across every session, on your own hardware. Key features: . Semantic search (find memories by meaning, not keywords) . Per-agent API keys with granular permissions . 3D constellation viewer with live agent activity . File vault (attach real documents to memories) . CSV audit log (full transparency on every agent action) . Pre-Strata history import (your memory doesn't start at install day) . Global MCP kill switch (emergency brake, only a human can undo) . Automatic deduplication . 10 structured thought types . Backend using PostgreSQL . Runs on a Raspberry Pi Always on, always local, always yours.

Persistent codebase intelligence that gives agents memory across sessions.

This is a cross-platform desktop application that allows you to chat with locally hosted LLMs and enjoy features like MCP support

Local-first AI agent that unifies 25+ plugins (Spotify, Gmail, GitHub, Notion, Telegram, Vercel, and more) into a single orchestration framework. AES-256-GCM encrypted credentials, zero telemetry, supports Claude, GPT-4o, Gemini, and Ollama.

Persistent memory for any AI assistant. Zero token cost until recall. Stores memories locally in SQLite, ranks by 6-factor scoring, returns results in TOON compact format (79% smaller than JSON). 17 MCP tools, 20 HTTP endpoints, 25 CLI commands. 4 tiers from keyword to autonomous with local LLMs via Ollama. 97.8% recall accuracy on ICLR 2025 LongMemEval benchmark. Works with Claude, ChatGPT, Grok, Cursor, Windsurf, Continue.dev, OpenClaw, Llama, and any MCP client. Open source, MIT licensed, written in Rust.

Code indexing MCP server to provide context to coding agents.

Run & Manage Long Processes - install, scrap, encode video, compile Search & Read & Edit files - including diff editing for speed and precision

Local-first MCP code intelligence server with 37 tools — hybrid search, blast-radius analysis, diff-aware risk-scored review, and bi-temporal memory pinned to git SHAs. Stops AI agents from hallucinating function names that don't exist in your codebase. Published 60-task retrieval benchmark with reproducible harness. MIT.

The graph augmented, smart and local-first Personal Context Manager, or Personal Agent Memory, just works. Nowledge Mem ingests your agent’s summarized memories through its MCP tools or prompts, enabling you to enhance your memory and knowledge using a GraphRAG approach. By leveraging LLMs, embedding models, BM25, graph algorithms, and a built-in knowledge graph, Nowledge Mem makes building, searching, and exploring your personal knowledge easier than ever.

Native macOS MCP server. Gives AI assistants access to Mail, Calendar, Teams, OneDrive, Contacts, Notes, Reminders, Finder, Word, Excel, PowerPoint — 82 tools, all running locally. No cloud, no API keys. GDPR compliant by architecture.

Local-first knowledge base MCP server. Let AI agents (Claude, Cursor, Gemini CLI, etc.) read and write your Markdown notes via MCP protocol. Supports Streamable HTTP and stdio transports.

Memory with integrity for AI coding agents. SAME tracks provenance, flags stale knowledge, and surfaces contradictions; so your AI trusts what's current, not what's outdated. 17 MCP tools: semantic search, cross-vault federation, session handoffs, decision logging, knowledge graph, trust-aware retrieval, consolidation, health analysis. Provenance on every write. Stale notes rank lower automatically. SQLite + vector search. Ollama, OpenAI, LM Studio, or keyword-only. One 12MB Go binary. No cloud, no API keys, no telemetry. Your notes never leave your machine. Claude Code, Cursor, Windsurf, Codex CLI, Gemini CLI, and any MCP client.

Edit Word (.docx) documents with tracked changes, redlines, and formatting preservation. Built for AI coding agents. Surgical text replacement, comments, footnotes, and document comparison — all local, no cloud upload. MIT licensed.

Tells you how you and your Claude agent actually work together. Reads your CLAUDE.md, hooks, skills, memory, and scheduled tasks — then writes you a letter. Three core tools: collab (collaboration scoring + lint), security (secrets scan, prompt-injection surfaces, rule conflicts), and health (orphan jobs, overlap, structural drift). Local-only — no API keys, no data leaves your machine. Source: github.com/bleedmode/dearuser

Local-first memory firewall and MCP server for AI agents. Audrey gives coding agents recall, memory capsules, preflight checks, reflexes, validation, tool-trace learning, contradiction handling, and SQLite/sqlite-vec storage.

terminal-history-mcp gives AI agents (Claude Code, Cursor, Cline, Zed, any MCP client) full-text search over your shell history. - Local-only, no cloud — SQLite FTS5 at ~/.terminal-history-mcp/history.db - Supports zsh, bash, fish - Captures cwd, exit code, duration via opt-in shell hook - 11-pattern secret redaction runs BEFORE insert (GitHub PATs, OpenAI keys, AWS, JWTs, bearer tokens, env vars, CLI flags, basic-auth URLs) - Hash-deduped, idempotent reindex - WAL mode — safe across parallel sessions Ask: "When did I last ssh into staging?", "Show failed commands today", "What did I run in /etc/nginx yesterday?", "Show context around kubectl apply".

peek records the user's actual logged-in browser (DOM via rrweb, console events, network metadata, optional response bodies via opt-in Deep capture) through a Chrome MV3 extension. The extension ships events through a native-messaging stdio bridge to a local MCP server (peek-mcp), which persists them to a SQLite database at ~/.peek/sessions.db. AI coding agents (Claude Code, Cursor, Cline, Windsurf) read sessions from the database via 10 MCP tools: Tool What it does list_recent_sessions List recently recorded sessions (id, origin, ts, event count). get_session_summary LLM-readable narrative summary of a session. get_session_console_errors Console errors recorded in a session. get_session_network_errors Failed/notable network requests in a session. get_user_action_before_error Last N user actions before a console error. generate_playwright_repro Generate a runnable Playwright test from a session. get_dom_snapshot Reconstruct the DOM at a given timestamp. query_dom_history Timeline of attribute/text changes for a selector. request_authorization Side-panel consent for write actions (Level 3). execute_action Dispatch a UI action (gated by permission level + destructive blocklist). Why local-first matters Every other "browser session for AI" tool ships to a vendor cloud. peek's SQLite + extension live on the user's machine — no remote endpoints, no telemetry. The privacy policy (docs/peek/PRIVACY_POLICY.md) is the source of truth. Install # 1. Add the MCP server to Claude Code claude mcp add peek -- npx -y @peekdev/mcp # 2. Install the Chrome extension from the Chrome Web Store # (link added once the CWS listing is approved)