About Yggdrasil
Yggdrasil is a local persistent-memory MCP server for coding agents, aimed at developers who want project decisions retained across sessions and tools. It stores memories in SQLite and Markdown, provides keyword search, and optionally adds semantic search and background consolidation.
Key features
- 01Share persistent memory across sessions
- 02Search keywords with SQLite and FTS5
- 03Add optional local-model semantic search
- 04Store memories as Markdown
- 05Expose memory operations through MCP tools
- 06Sync directly between linked devices
Use cases
- 01Carry project context into new coding sessions
- 02Recall relevant experience from other projects
- 03Import memories from transcripts and Obsidian
- 04Sync local memory across connected devices
MCP support
Yggdrasil exposes tools for searching, recalling, and saving memory through an MCP server. Its installer can register these tools with detected agent hosts.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
uvx yggdrasil-memoryYGG_ENGINE_URLBase URL of the Yggdrasil engine HTTP API.YGG_ENGINE_TOKENSecretBearer token for authenticating to the Yggdrasil engine (from ~/.yggdrasil/token after `ygg install`).Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.
Tools
6Saved tool list · updated 2w ago
This tool takes no parameters.
Details
From the project README🌳 Yggdrasil
Stop re-explaining your project to every new AI session.
One local memory for Claude Code, Codex, and every MCP agent — shared across sessions, tools, and projects. Zero dependencies. Nothing leaves your machine.
Install · How it works · Numbers · Compare · FAQ
Read this in: Русский · 简体中文 · Español · Français · 日本語 · Deutsch
Every new chat, your AI forgets. You re-explain the project, the decisions, the gotchas — every time, in every tool. Yggdrasil is a tiny always-on memory that any agent plugs into. Open a new session, in any project, with any AI, and it already knows what you decided, what broke, and what's still open.
$ cd ~/projects/checkout-api && claude # a brand-new session
🌳 Yggdrasil (injected automatically at session start)
• [project_status] payments refactor: idempotency keys added; open: e2e tests
• [lesson] webhook 401 → signing secret rotated; update env + redeploy
> "have I solved a flaky websocket reconnect anywhere before?"
🌳 recall → found in project `realtime-dash`:
refresh the token *before* opening the socket, then retry with capped backoff.
No "let me remind you what we did yesterday." It's just there.
🚀 Install
Two commands, inside Claude Code (the plugin launches via uv):
/plugin marketplace add VonderVuflya/Yggdrasil
/plugin install yggdrasil
The engine lazy-starts on first use and generates its own local token — no API key, no cloud, nothing to configure. Codex and Cursor use the same flow.
All other channels — CLI daemon, Homebrew, npm, Claude Desktop, from source…
| Host / tool | Command |
|---|---|
| uvx (recommended CLI) | uvx --from yggdrasil-memory ygg install |
| npm / npx | npx yggdrasil-memory install |
| pipx | pipx install yggdrasil-memory && ygg install |
| pip | pip install yggdrasil-memory && ygg install |
| Homebrew (macOS) | brew install VonderVuflya/tap/yggdrasil && ygg install |
| Claude Desktop (app) | drag the .mcpb from the latest release onto Settings → Extensions, paste your token (ygg token) — the desktop app then shares the same memory as your CLI agents (guide) |
| from source | uvx --from git+https://github.com/VonderVuflya/yggdrasil.git ygg install |
ygg install is a one-time guided setup: it installs an always-on background service, registers the MCP tools with every agent host it finds — Claude Code, Codex, OpenCode — and, if your hardware allows, recommends optional local models (or pick none to stay zero-config).
OpenCode — nothing to configure
Install OpenCode first, then run ygg install (or ygg redeploy if Yggdrasil is already set up) — the entry is written for you and merged into any existing opencode.json. Confirm with:
opencode mcp list # -> ✓ yggdrasil connected
Installed OpenCode after Yggdrasil? Just re-run ygg install.
If you'd rather write it by hand, note that OpenCode's schema differs from Claude's in four places at once — servers live under mcp (not mcpServers), type is required, command is one array (not command + args), and env is environment (not env), so the Claude snippet won't port:
// ~/.config/opencode/opencode.json
{
"$schema": "https://opencode.ai/config.json",
"mcp": {
"yggdrasil": {
"type": "local",
"command": ["/path/to/python3", "~/.yggdrasil/scripts/ygg_mcp_server.py"],
"enabled": true,
"environment": { "YGG_ENGINE_URL": "http://127.0.0.1:42069" }
}
}
}
No token goes in the config — the engine reads the 0600 ~/.yggdrasil/token itself. Run ygg doctor if the tools don't show up.
There is also a yggdrasil-memory skill for any Claude surface: MCP connects the tools, the skill teaches the agent when to use them. Use both for the best behavior.
Try it with nothing installed and a throwaway DB: uvx --from yggdrasil-memory ygg serve --reset --db /tmp/ygg.sqlite.
Then just work: ask your agent "recall what we decided about this project", tell it "remember this decision" — next session it's already there. Verify the install any time with ygg doctor.
Already have history? Seed memory from your existing Claude Code + Codex transcripts, Obsidian vaults, and CLAUDE.md repos — distilled locally:
ygg seed --dry-run # see what it would import; drop the flag to distill for real
Leaving another memory tool? ygg import --from mcp-memory --path memory.json pulls its whole store into Yggdrasil (deduped, secret-guarded) — then you can delete it.
Why
- 🧠 Persistent — decisions, lessons, and project status survive across sessions.
- 🔌 One brain, every tool — Claude Code, Codex, OpenCode, and any MCP host share the same memory.
- 🌐 Cross-project recall — "this looks like what you did in project B — reuse it?"
- 🧹 Curated, not captured — your agent saves the few things that matter; governance dedupes and archives, never deletes.
- 🌱 Self-maintaining (opt-in) — a small local model consolidates memory in the background. Zero API tokens.
- 🪪 One identity everywhere — an optional name and persona every agent picks up, so Claude Code and Codex feel like the same assistant.
- 🔒 100% local — your memory lives on your machine. No cloud, no account, no telemetry.
🧠 How it works
Yggdrasil is memory + tools — the intelligence is your LLM. It just makes sure the right memory is in front of the right agent at the right moment.
- 🛎️ Always-on daemon — a tiny local service (~21 MB RAM) your agents reach over MCP tools (
ygg_search,ygg_recall,ygg_remember…). - 🪝 Hooks — session start auto-injects identity, project status, and open follow-ups (~300 tokens); an optional per-prompt hook auto-recalls memory relevant to each request.
- 📌 Ranking — pinned and frequently-recalled memories surface first.
- 🧹 Governance — duplicates and conflicts are queued for review; changes are non-destructive (archive, never delete).
- 📓 Obsidian — every memory doubles as a plain-Markdown note you can read, edit, and grep.
🎛️ Memory tiers — zero-config by default
Out of the box, Yggdrasil runs on SQLite + FTS5 with zero dependencies — instant keyword search, no models, nothing to download. Optional local models add two independent tiers:
| Tier | You add | You gain |
|---|---|---|
| 0 · default | nothing — SQLite + FTS5 | keyword search, zero deps, instant — recall@1 = 0.77 |
| 1 · semantic | an embedding model (all-minilm 45 MB · paraphrase-multilingual ~560 MB) | search by meaning, across languages — recall@1 = 0.94, recall@3 1.00 |
| 2 · self-maintaining | a small LLM (qwen2.5:1.5b ~1 GB) | background dedupe/merge of memory (propose-only) |
The runtime only computes vectors and runs the background model — every memory and every vector stays in the same local SQLite.
ygg install scans for Ollama, LM Studio and llama.cpp, offers to start whichever is installed but idle, and then shows one menu per job with green for models you already have and red for models it would download. Pick a row, and it writes embed_backend, embed_url and distill_url for you — including the part nobody guesses right, that embed_url wants the /v1 base while distill_url wants the host root. A runtime on another machine is one URL: paste it and the dialect is detected. ygg recommend shows the same scan plus the full catalog without changing anything.
Full model menu
Embeddings (semantic search):
| Model | Size | Good for |
|---|---|---|
all-minilm | 45 MB | English, tiny & fast |
nomic-embed-text | 274 MB | English, better quality (768d) |
mxbai-embed-large | 670 MB | English, high quality (1024d) |
paraphrase-multilingual | ~560 MB | multilingual (EN/RU + 50 langs, 768d) |
bge-m3 | 1.2 GB | multilingual, top quality (heavier) |
Embedding backend — Ollama by default. ygg install sets all of this for you
once you pick a runtime; the manual route below is for scripted setups and for
changing one thing later. To use an OpenAI-compatible /v1/embeddings server
instead (llama.cpp's llama-server --embeddings, OpenRouter, LM Studio, vLLM),
set embed_backend:
# local llama.cpp — no key needed
ygg config set embed_backend openai
ygg config set embed_url http://127.0.0.1:8080/v1
ygg config set embed_model bge-small-en-v1.5
ygg redeploy
# OpenRouter — free embeddings, no GPU needed
ygg config set embed_backend openai
ygg config set embed_url https://openrouter.ai/api/v1
ygg config set embed_model nvidia/llama-nemotron-embed-vl-1b-v2:free
ygg config set embed_api_key sk-or-... # or export YGG_EMBED_API_KEY
ygg redeploy
The key is stored in ~/.yggdrasil/embed_api_key (0600) rather than
config.json, and reaches the daemon as a file path — so it never shows up
in ps, the launchd plist or the systemd unit. ygg config list masks it.
Check it took with ygg doctor — dense should name your model:
✓ dense active (nvidia/llama-nemotron-embed-vl-1b-v2:free)
Wiring LM Studio or OpenRouter by hand? The gotchas (two URL shapes, model ids, JIT loading, inference vs provisioning keys, privacy filters) are in docs/model-backends.md.
Staying local still wins on quality and privacy: on the 232-memory / 110-query
corpus, local paraphrase-multilingual scores recall@1 0.964 vs 0.946
for the free hosted model — and your memories never leave the machine. Reach for
a hosted backend when the box can't run Ollama, not to chase accuracy.
Bigger vectors do not buy accuracy here — on the same corpus
mxbai-embed-large (1024d) scores 0.809 and nomic-embed-text (768d) 0.818,
a difference their confidence intervals swallow whole. What actually moves the
number is whether the model handles your languages: both are English-only and
collapse to 0.40–0.45 on cross-language queries, where the multilingual default
holds 0.95.
Background consolidation (small LLM):
| Model | Size | Good for |
|---|---|---|
qwen2.5:0.5b | ~400 MB | tiny, fast on CPU |
qwen2.5:1.5b | ~1 GB | best CPU default |
llama3.2:3b | ~2 GB | better quality, slower on CPU |
The engine itself is swappable — any service meeting the MemoryBackend contract is a drop-in (YGG_ENGINE_URL); see docs/backend-boundary.md.
🔗 One memory across your machines
Work on a laptop and a beefier box? Link them once and they stop being two brains.
# on the laptop
$ ygg link
🔗 pairing code valid 5 minutes · single use
ygg://192.168.3.152:42070/a4e0df489737#FOucjgJ1FrUB…
# on the other machine — paste it
$ ygg link ygg://192.168.3.152:42070/a4e0df489737#FOucjgJ1FrUB…
✓ linked with laptop https://192.168.3.152:42070
✓ laptop: fetch 84 · push 12 · delete 0 · tombstones 0
That's it. From then on a memory written on one machine is on the other in about
a second, ygg delete removes it from both, and either machine keeps working
with its own local copy when the other is asleep — reads never wait for the
network. When a machine comes back, the two compare fingerprints and repair
whatever they missed, without you running anything.
There is no polling. Yggdrasil is called a handful of times a day, so writes push as they happen and a sync is triggered by reads; idle machines send nothing.
What this is not: a cloud. There is no relay, no account, no server of ours in
the path. The machines talk to each other directly over your own network, over
TLS with each other's certificate pinned, using a key pair that exists only on
those two machines. The pairing code itself names the issuing machine's
certificate, so even the very first connection is verified — something else on
the network answering in its place gets refused before the code's secret is sent. Nothing binds a network port until you run
ygg link — install Yggdrasil and it stays on loopback, exactly as before.
ygg link --list # what's linked, and whether writes are actually arriving
ygg link --sync # reconcile now, don't wait for a read
ygg unlink <name> # revoke that machine's key here
Details: what travels, what stays, and when to use ygg sync instead
Travels: memories, their relations, and deletions. Stays put: how often you opened a memory and which machine you read it on — that's local ranking, not knowledge. Embedding vectors normally stay local too, with one exception: if both machines run the same embedding model, the vector rides along, so a compute box can do the embedding for the laptop and the laptop never runs a model.
The newer edit wins. Edit a memory on both machines before they talk and the more recent edit is kept on both (a shorter correction included).
Deletions win. If you delete a memory on one machine and edit it on the other
at the same time, the deletion wins and the edit is lost. That's deliberate:
"I deleted the leaked token" has to hold even when two clocks disagree. When you
mean changed my mind, archive it instead — nothing is lost, and
ygg unarchive brings it back on both machines (the later decision wins).
A firewall is caught at pairing. The machine that printed the code calls the
other one back straight away; if it can't get through, ygg link says so and
prints the one command that opens the port to that machine only (ufw, firewalld,
Windows). ygg link --list and ygg doctor keep checking afterwards: a peer
shows ✓ only once writes are confirmed delivered, otherwise the error and how
many changes are waiting. A machine that gets a new address from DHCP is found
again the next time it makes contact.
ygg sync (git) is still there and does a different job: machines that can't
see each other, a work laptop behind a different network, a bare repo on a USB
stick you carry. Use ygg link for machines on one network, ygg sync for
everything else. Both share the same merge rules, so you can run both. If both
machines are on a Tailscale tailnet, ygg link --tailscale works across networks too.
If the pairing code shows the wrong address (it prefers real Ethernet and
Wi-Fi over VM bridges and VPNs, but can still guess wrong), pin it:
ygg config set sync_host 192.168.3.152 — or print the code on the other machine:
the side that pastes it always uses the address the route actually takes.
📊 The numbers
Measured by eval/ygg_eval.py — 232 memories, 110 labelled queries, ranking weights tuned on the dev split only, so holdout is the unbiased number (recall@1, with the paraphrase-multilingual model):
| Search view | holdout recall@1 | recall@3 | zero-dep lexical |
|---|---|---|---|
| Within a project (the real path, pool ~11) | 0.94 | 1.00 | 0.76 |
| Whole store (no filter, pool 232) | 0.72 | 0.87 | 0.69 |
Within a project — the path you use — the right memory is #1 for 0.94 of queries and in the top 3 every time (recall@3 = 1.00). Searching the whole store with no filter is harder (recall@1 0.72, recall@3 0.87 across all 232). Zero-dep lexical mode already solves keyword and code-identifier queries (1.00); the local model adds meaning and cross-language (crosslingual 0.25 → 0.95). The full breakdown in BENCHMARKS.md has 95% CIs, pool sizes, and per-class scores — rerun it in a minute: python3 eval/ygg_eval.py --report.
Corpus size is part of the number. On a real 4,799-memory store the default multilingual embedder falls to the lexical baseline (recall@1 0.55); bge-m3 holds 0.775 there. If your store grows into the thousands, pick bge-m3 — details in BENCHMARKS.md.
🆚 Yggdrasil vs the rest
Everyone else either auto-captures transcripts or sells you a cloud. Yggdrasil's bet: keep the few things that matter, curated and de-duped, in plain rows you own — and share them across every tool and project.
| Yggdrasil | Built-in memory (Claude Code · Codex) | claude-mem | mem0 / OpenMemory | basic-memory | |
|---|---|---|---|---|---|
| Curated decisions / lessons / status (not transcripts) | ✅ | ⚠️ auto-notes | ❌ captures everything | ⚠️ | ⚠️ free-form notes |
| One memory across tools | ✅ | ❌ vendor-siloed | ✅ | ✅ | ✅ |
| Cross-project recall ("solved this in project B") | ✅ | ❌ repo-scoped | ⚠️ | ⚠️ | ⚠️ |
| 100% local by default | ✅ | ✅ | ⚠️ cloud sync add-on | ❌ hosted-first | ✅ |
| Zero dependencies (stdlib + SQLite) | ✅ | — | ❌ Node + Bun + worker daemon | ❌ Docker + Qdrant + LLM key | ❌ |
| Works with no LLM & no API key | ✅ | ✅ | ❌ AI-compresses | ❌ | ✅ |
| Semantic search, fully local | ✅ opt-in Ollama | ❌ grep-only | ⚠️ optional Chroma | ⚠️ needs API key or Docker stack | ❌ |
| Plain Markdown you own (Obsidian-ready) | ✅ | ✅ | ❌ | ❌ | ✅ |
Closest neighbor — claude-mem: capture-everything memory that records and AI-compresses every session (Node 20+ and Bun, a persistent worker daemon; Chroma optional). Yggdrasil is the opposite bet: a small, high-signal store instead of a growing firehose. mem0 is an SDK plus a hosted platform for building apps that remember their users — even self-hosted it needs an LLM API key. Built-in memories are genuinely useful — and structurally siloed: one vendor, one repo, one machine, literal grep. Yggdrasil is the layer above them (and ygg seed can bootstrap itself from those same transcripts). Different layer entirely: context-mode (live context window) and Context7 (fresh library docs) — both pair fine with Yggdrasil.
🧰 Commands
Agents see six MCP tools: ygg_health, ygg_bootstrap, ygg_search, ygg_recall, ygg_remember, ygg_materialize — auto-registered by the plugin or ygg install.
Full ygg CLI reference
Memory ops
| Command | What it does |
|---|---|
ygg recall --query "…" | Cross-project search — "have I done this anywhere?" |
ygg search --project P --query "…" | Project-scoped search (--type, --tag, --limit, --json) |
ygg remember --project P --type lesson --content "…" | Save a durable memory (secret-guarded, deduped) |
ygg bootstrap --project P | Pull a project's memory before starting work |
ygg pin --id ID · ygg unpin --id ID | Pin a memory so it reliably surfaces |
ygg relate --from A --rel solves --to B · ygg relations --id ID | Link memories (solves/supersedes/contradicts) · see why a memory exists / what replaced it |
ygg supersede --id OLD --by NEW | Archive an outdated memory — --by records what replaced it |
ygg unarchive --id ID | Bring an archived memory back (on linked machines too) |
ygg materialize --id ID --project P | Export one memory to an Obsidian note |
ygg export-native --project P | Write a curated digest into AGENTS.md/MEMORY.md — feed Claude Code & Codex's native memory |
ygg import --from TOOL --path P | Migrate another memory tool's store into Yggdrasil (mcp-memory, basic-memory; --dry-run first) |
ygg review [--apply] | Work the governance queue — consolidate duplicates, flag stale/conflicting memories (archive-only, reversible) |
ygg delete --id ID · ygg reset … | Hard-delete one memory · bulk-undo a bad seed (confirms first) |
Cold start
| Command | What it does |
|---|---|
ygg seed | Distill Claude Code + Codex transcripts, Obsidian vaults, CLAUDE.md repos — incremental, deduped, fully local |
ygg seed --dry-run · --force | Discover + estimate only · re-distill everything |
ygg seed --schedule 03:30 | Nightly auto-distill (launchd) — memory keeps itself fresh; off / status |
ygg sync --repo <your-git-repo> | Sync memory across machines through your own git repo — plain JSON files, no cloud in the loop |
ygg link · ygg link <code> | Link two of your machines directly on one network — live shared memory, pinned TLS, no relay |
ygg link --list · ygg unlink <name> | Show linked machines · revoke one |
ygg distill --source PATH | Distill one dir/file into lessons |
ygg reindex | Backfill missing embeddings (restores dense recall) |
Service & setup
| Command | What it does |
|---|---|
ygg install · ygg doctor · ygg update | Guided setup · diagnose with actionable fixes · upgrade |
ygg config | Show/set persistent settings (list · get · set · unset) |
ygg status · start · stop · restart · logs | Manage the always-on daemon |
ygg hooks · unhooks · register | SessionStart hook on/off · (re)register MCP |
ygg recommend · token · uninstall | Model catalog · print auth token · remove everything |
Give it a personality — edit ~/.yggdrasil/identity.json:
{ "name": "Jarvis", "persona": "concise, proactive, dry wit", "user_facts": ["prefers TypeScript", "ships small PRs"] }
Heavy seeding, weak laptop? Point distillation at any box on your LAN — a desktop with Ollama, LM Studio, llama.cpp, even an iPhone running a local-LLM server app: ygg config set distill_url http://<box>:11434. Yggdrasil auto-detects the API dialect (Ollama or OpenAI-compatible); your data still never leaves your network — details in docs/ygg-cli.md.
❓ FAQ
Claude Code already has built-in memory — why Yggdrasil?
Built-in memories are per-vendor, per-repo, per-machine, and retrieved by literal text match. Yggdrasil is the layer above: the same memory in Claude Code, Codex, and any MCP host, recall across projects, optional semantic search — still 100% local. It bridges them both ways: ygg seed distills your existing native memory + transcripts into the shared brain, and ygg export-native writes a curated digest back into AGENTS.md/MEMORY.md — so even a fresh clone or a tool without Yggdrasil still gets your curated memory.
Does it send my code or memory to the cloud?
No. The engine, the database, and the optional models all run locally. No account, no telemetry. The only outbound call is a version check against PyPI.
If you link two of your own machines with ygg link, memories travel between those two machines over your own network — directly, over TLS with a pinned certificate, with no relay and no server of ours in the path. Until you run that command nothing listens beyond loopback.
Can a malicious repo plant instructions in my memory?
Recalled memory is injected as clearly-framed reference notes, not instructions, with where each note came from and a marker when it belongs to a different project — so the agent treats it as data and asks before acting on anything in it. Secrets are redacted from transcripts before distillation and refused on write. Still: memory is only as trustworthy as what wrote it — review ygg review now and then.
Does it automatically remember everything?
No — by design. Retrieval is automatic; writing is deliberate (the agent calls ygg_remember for durable lessons). Capture-everything pollutes memory and burns tokens, so we don't. The optional background model consolidates what's already saved (propose-only).
Do I need a GPU or an API key?
No. The default is pure lexical search — zero dependencies, instant. Semantic search is opt-in and uses a local model via Ollama. The installer recommends one that fits your hardware.
How heavy is it, and what does it cost in tokens?
The engine idles at ~21 MB RAM (lexical default) with ~0% CPU; disk is tens of KB per memory. Session start injects ~300 tokens; each tool call returns a small snippet. All heavy work (indexing, embeddings, consolidation) runs off-LLM on your machine.
Can I edit or delete memories by hand?
Yes. Memories materialize to Markdown notes in an Obsidian vault — read, edit, or remove them like any file. Governance and review never delete anything — they archive, which is reversible. Only you hard-delete, explicitly, with ygg delete (and on linked machines that delete reaches both).
🚦 Status & roadmap
Alpha. The happy path and the governance loop are gate-tested (scripts/run_gates.sh); not yet hardened for multi-user or production use. macOS today; Linux/Windows service installers are built and in final on-device testing.
Shipped recently: 🔗 machine-to-machine memory (ygg link) · relation graph (ygg relate: solves / supersedes / contradicts).
Next: 🛰️ cross-surface (one memory across CLI, web, and phone — the HTTP facade is in, see docs/cross-surface.md) · 🐧 Linux/Windows GA.
🤝 Contributing
Issues and PRs welcome. Run scripts/run_gates.sh and python3 -m unittest discover -s tests before submitting — all gates must stay green.
📜 License
GNU AGPL v3.0 — see LICENSE. Free and open source: use, modify, self-host, redistribute. If you modify it or offer it as a network service, you must release your source under the same license.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.