Agenten stärken
Der eigene Befehl des Herausgebers. Kein PlugMCP-Endpunkt.
Host a copy with our keys and logs.
npx -y [email protected]Claude Code hat keinen Installationslink für einen lokalen Befehl. Kopieren Sie den Befehl unten und führen Sie ihn in einem Terminal aus.
OPENAI_API_KEY=<secret> NAN_FORGET_EMBEDDING_PROVIDER=<value> NAN_FORGET_USER_ID=<value> NAN_FORGET_PROJECT_ROOT=<value> claude mcp add io.github.nanmesh-nan-forget -- npx -y nan-forget@0.6.8In einem Terminal ausführen.
{ "mcpServers": { "io.github.nanmesh-nan-forget": { "command": "npx", "args": [ "-y", "[email protected]" ], "env": { "OPENAI_API_KEY": "<secret>", "NAN_FORGET_EMBEDDING_PROVIDER": "<value>", "NAN_FORGET_USER_ID": "<value>", "NAN_FORGET_PROJECT_ROOT": "<value>" } } }}claude_desktop_config.json
Remote-URL oder ein Befehl auf Ihrem Rechner? Worin sich die beiden Installationen unterscheiden
Auf PlugMCP ausführen
A new server you create. It is not this listing's official endpoint. Use the publisher install when you want their URL and their credentials.
StartenVom Herausgeber
Long-term memory for AI coding tools.
Your AI forgets everything when the session ends. NaN Forget fixes that.
npx nan-forget setup
That's it. The wizard installs Ollama, embeddings, Claude hooks, MCP config, and a project for Codex-style agents. Restart Claude Code or reopen Codex. Your AI now remembers.
AI that remembers. Local-first cross-session, cross-project long-term memory for coding tools.
AGENTS.mdNo API keys needed. No Docker needed. Runs locally. Free forever.
flowchart LR
A["You talk to your AI tool"] --> B["It learns things"]
B --> C["nan-forget saves to SQLite"]
C --> D["Session ends"]
D --> E["New session starts"]
E --> F["nan-forget loads context"]
F --> G["Your AI remembers"]
~/.nan-forget/memories.db. Aging memories get automatically compacted into long-term entries.You never call save or search manually. Here's what happens behind the scenes:
Four layers of protection ensure nothing is lost:
Codex and similar agents work well with nan-forget, but they usually need instruction files or shell/API fallbacks instead of Claude's hook model:
npx nan-forget setupAGENTS.md in your repo. It tells Codex-style agents to sync, search, save, and checkpoint automatically.localhost:3456 or local commands like nan-forget sync, nan-forget search, nan-forget add, and nan-forget checkpoint.The REST API and CLI now mirror the important memory workflows closely enough that memories saved by Claude are searchable from Codex and vice versa.
# Start the REST API
nan-forget api
# The system prompt tells your agent exactly what endpoints to call
nan-forget prompt
Type these in Claude Code:
Claude uses MCP. Codex can use AGENTS.md plus CLI/REST fallback:
# Start the API
nan-forget api
# Get the system prompt for your agent
nan-forget prompt
Codex, Cursor, and Claude all share the same memory database.
curl http://localhost:3456/memories/search?q=auth
curl -X POST http://localhost:3456/memories/sync -d '{"project":"my-app"}'
curl -X POST http://localhost:3456/memories/checkpoint \
-H 'content-type: application/json' \
-d '{"task_summary":"Fixed auth regression","problem":"Expired tokens were not refreshed","solution":"Added refresh handling in middleware","files":["src/auth.ts"],"concepts":["auth","jwt"],"project":"my-app"}'
nan-forget add "We use FastAPI, not Django. Railway deploys faster."
nan-forget add --type decision "Auth is Clerk, not custom JWT"
nan-forget search "what auth system"
nan-forget stats
Everything below is for developers who want to understand how nan-forget works under the hood.
LLMs have no memory between sessions. Every conversation starts from zero. You re-explain your stack, Claude contradicts decisions from last month, and context disappears when the session ends.
Existing solutions (Mem0) target app developers embedding memory into products. We target you — the developer using AI tools daily who wants AI that just remembers.
flowchart TB
subgraph Short["Short-Term Memory"]
MD[".md files<br/>Current session context<br/>Disposable scratch paper"]
end
subgraph Long["Long-Term Memory"]
DB["SQLite + sqlite-vec<br/>~/.nan-forget/memories.db<br/>Semantic search + decay"]
end
subgraph Auto["Automatic Processes"]
Hook["Hooks (3)<br/>PostToolUse: .md → DB<br/>UserPromptSubmit: auto-recall<br/>SessionEnd: transcript sweep"]
Consolidate["Consolidation Engine<br/>Clusters + summarizes aging memories"]
GC["Garbage Collection<br/>Decay, dedup, expiry"]
end
MD -->|"hook intercepts"| Hook
Hook -->|"nan-forget add"| DB
DB --> Consolidate
Consolidate --> DB
DB --> GC
Short-term memory = Claude's built-in .md files. Disposable. Current session only.
Long-term memory = SQLite database with sqlite-vec vector search. Single file. Permanent. Searchable across all sessions, all projects, all LLM tools.
Three hooks handle memory automatically:
.md file writes and saves them to the DB.nan-forget recall on every user message, auto-searching memory for relevant context.Memory search follows the same path as human recall:
flowchart LR
Q["Query"] --> S1["Stage 1: Recognition<br/>Fast vector match<br/>Returns summaries only"]
S1 --> S2["Stage 2: Recall<br/>Full content fetch<br/>Cross-project expansion"]
S2 --> S3["Stage 3: Association<br/>Spreading activation<br/>Related memories surface"]
S3 --> R["Results ranked by<br/>similarity x decay x frequency x confidence"]
Scoring formula:
final_score = vector_similarity * decay_weight * frequency_boost * confidence_boost
decay_weight = (0.5 ^ (days / 30)) ^ (1 - confidence)
frequency_boost = log2(access_count + 1) / 10 + 1
confidence_boost = 0.5 + 0.5 * confidence
High-confidence memories (debate-validated, human-approved) decay much slower and rank higher. A core memory at 0.85 confidence decays at ~15% of normal rate — effectively permanent unless superseded. Cross-project search means auth decisions from Project A surface when you work on Project B.
Aging memories don't just get deleted — they get compacted into long-term entries:
flowchart TB
A["10+ aging memories<br/>about the same topic"] --> B["Cluster by project + type<br/>+ vector similarity > 0.8"]
B --> C{"OpenAI key<br/>available?"}
C -->|"Yes"| D["LLM summarizes cluster<br/>into 1-2 sentences"]
C -->|"No"| E["Deterministic merge<br/>concatenate + deduplicate"]
D --> F["Save consolidated entry<br/>with fresh vector embedding"]
E --> F
F --> G["Archive originals<br/>with backlink"]
Triggers automatically after every 10 saves or 24 hours. No user action needed.
memory_save accepts structured fields for richer vector representation:
All fields are embedded together into a single vector. Searches for "JWT auth bug" find memories tagged with those concepts even if the content text doesn't match literally.
Not all memories are equal. Debate-validated and human-approved memories are core — they decay slower, rank higher in search, and survive garbage collection longer.
Core memories are designed for the upcoming nan-debate system — multi-AI debate results validated by human approval get persisted as high-trust knowledge that almost never fades.
After completing a task, call memory_checkpoint with task_summary, problem, solution, files, concepts, and project. Saves the full problem→solution context to long-term memory. Every completed task = one checkpoint.
memory_compress scans .claude/projects/*/memory/ for .md files already persisted to the DB. Persisted files are replaced with minimal stubs. Reduces context window bloat.
Shares the same SQLite database as the MCP server — memories saved by Claude are searchable from Codex and vice versa.
POST /memories — Save a memory (supports problem/solution/files/concepts)
POST /memories/checkpoint — Save completed-task context
POST /memories/sync — Lightweight session handshake
GET /memories/search?q=... — Semantic search
GET /memories/:id — Get by ID
PATCH /memories/:id — Update
DELETE /memories/:id — Archive
POST /memories/consolidate — Compact aging memories
POST /memories/clean — Garbage collection
GET /memories/stats — Memory health
GET /memories/instructions — System prompt for LLMs
Get the system prompt for any LLM:
nan-forget prompt
# or
curl http://localhost:3456/memories/instructions
Auto-detection: Ollama running? Use it. Not running? Check for OPENAI_API_KEY. No config needed.
All data lives in a single SQLite file at ~/.nan-forget/memories.db. No Docker, no services, no data loss on updates.
user_id, status, project, typenan-forget export dumps all memories as JSON.flowchart TB
A["New memory saved"] --> B["Active in SQLite<br/>Searchable, scored"]
B --> C{"Accessed<br/>recently?"}
C -->|"Yes"| D["Score stays high<br/>frequency_boost increases"]
C -->|"No"| E["Decay weight drops<br/>0.5^(days/30)"]
E --> F{"Decay < 0.3?"}
F -->|"Yes"| G["Consolidation candidate<br/>Clustered + summarized"]
F -->|"No"| B
G --> H["New consolidated entry<br/>Originals archived"]
D --> B
E --> I{"Decay < 0.1?"}
I -->|"Yes"| J["Archived by GC"]
I -->|"No"| F
All cleanup is deterministic. No API calls. No LLM inference.
expires_at dateCore memories (confidence ≥ 0.85) survive GC far longer than regular ones because their decay formula dampens the time factor: decay^(1 - confidence). A 0.85-confidence memory at 200 days still has a decay weight above 0.1.
NaN Forget is built around three principles: lightweight, automatic, and local.
No Docker. No cloud services. No background processes eating RAM. The entire storage layer is a single SQLite file (~3 MB). Embeddings run through Ollama, which you likely already have. Memory operations (save, search, dedup, GC) use zero LLM calls — all deterministic.
Four hooks capture context at every stage of a session — you never call save manually:
.md file writes and persists themAging memories consolidate automatically. Duplicates merge. Unused memories decay on a 30-day half-life. No maintenance required.
Your data stays on your machine in ~/.nan-forget/memories.db. No accounts, no API keys required (Ollama is free and local), no telemetry. Backup is copying one file. Works across Claude Code (MCP), Codex/Cursor (REST API), and the terminal (CLI) — same database, same memories.
Most AI memory solutions (Mem0, claude-mem) are designed for app developers embedding memory into products, or require Docker/cloud services to run. NaN Forget is designed for you — the developer using AI tools daily who wants context that persists across sessions without managing infrastructure.
Key design differences:
problem, solution, concepts, and files fields — searches find related context even when keywords don't matchnpx nan-forget setup), no Docker, no containers, no services to managesrc/
sqlite.ts SQLite + sqlite-vec storage layer (schema, CRUD, vector search)
embeddings.ts OpenAI / Ollama abstraction
writer.ts Memory writer with dedup (>0.92 = merge)
retriever.ts Three-stage retrieval pipeline
consolidator.ts LLM summarization + deterministic fallback
cleaner.ts GC: decay, expiry, dedup, MEMORY.md sync
services.ts Service management (Ollama, REST API)
memory-md.ts MEMORY.md manager
types.ts Shared types (Memory, MemoryType, etc.)
mcp/server.ts MCP server, 13 tools
api/server.ts REST API server
cli/index.ts CLI commands + hook helpers
setup/index.ts Setup wizard (Ollama, hooks, MCP config)
.claude/
commands/nan-forget.md Slash command for manual control
hooks/memory-sync.js PostToolUse hook (auto-saves .md → SQLite)
hooks/session-end.js SessionEnd hook (transcript sweep for unsaved memories)
settings.json Hook config (PostToolUse + SessionEnd + UserPromptSubmit)
MIT License.
AI that remembers. Local-first cross-session, cross-project long-term memory for coding tools.
{ "mcpServers": { "io.github.nanmesh-nan-forget": { "command": "npx", "args": [ "-y", "[email protected]" ], "env": { "OPENAI_API_KEY": "<secret>", "NAN_FORGET_EMBEDDING_PROVIDER": "<value>", "NAN_FORGET_USER_ID": "<value>", "NAN_FORGET_PROJECT_ROOT": "<value>" } } }}Tools
Noch keine Tool-Liste. Remote-Server erhalten eine nach dem ersten Health-Check.
FAQ
Kopieren Sie auf dieser Seite die stdio-Konfiguration für Cursor, Claude Desktop, Claude Code oder VS Code. Sie nutzt die Installation des Herausgebers, keinen PlugMCP-Proxy.
Der Herausgeber nennt diese Umgebungsvariablen: OPENAI_API_KEY, NAN_FORGET_EMBEDDING_PROVIDER, NAN_FORGET_USER_ID, NAN_FORGET_PROJECT_ROOT.
Nan Forget läuft lokal über stdio.
Zuletzt geprüft: 2026-11-09. AI that remembers. Local-first cross-session, cross-project long-term memory for coding tools.