Itamos MCP Tools

Built in Greece, for the world. Free and open source under AGPL-3.0. Commercial licences for closed-source use.

---

The problem

Every LLM coding assistant faces the same wall: codebases are too large to fit in context. The standard response is to dump files — grep for something promising, cat it, hope for the best. At 50k files this breaks. At 240k files it never worked.

We built four tools that give an LLM structured perception of a codebase instead of raw file access, plus a sandboxed git to bring code in. The result is a model that navigates code the way a senior engineer does — starting from the architecture, narrowing to the module, reading only the segment it needs.

Read the full paper: docs/PAPER.md: how each tool works, why it was designed that way, and how the hosted sandbox runs, with diagrams.

---

Tools

Tool	What it does
master_architect	Project-aware code navigation. Builds a graph of your codebase, exposes topology, per-file bone structure, and segment-level addressing. The entry point for any codebase task.
read_file	Segment-addressed file editor. Returns a skeleton by default. Read segment N to get the code. Edit with verify then commit. Project-aware.
write_file	Create new files with automatic project placement. Refuses overwrites — edits go through read_file.
git	The original Itamos git tool, sandboxed for security: clone (remote URLs only, depth=1), status, commit. Everything stays inside the session workspace; push is not available.
web_skeleton	LLM-first web perception. 97% token reduction vs raw HTML. Actions: search, skeleton, read, click.

At least 97% fewer tokens than the usual shell workflow (grep, cat, run_cmd), measured against the best case where the model fixes the bug in one attempt. Real sessions save more, because the failed attempts common with raw shell access aren't counted.

---

How it works

Session model

Every client gets an isolated workspace. No shared state, no cross-session leakage. Workspaces are wiped after 10 minutes of inactivity.

master_architect navigation flow

The model is guided through four phases: scan to index the repo, topology to see the data-flow graph, bones to inspect a file, navigate or read_file to read the specific segment. The model never reads a file it has not first located in the graph.

read_file segment addressing

Files are parsed into named segments. A 5000-line file might have 40 segments. The model reads the skeleton, picks the segment it needs, reads that segment. Total context used is roughly 120 lines instead of 5000.

web_skeleton token reduction

Raw HTML of a modern web page runs 50,000 to 200,000 tokens. web_skeleton output is 500 to 3,000 tokens. The model reads the skeleton, picks the section id it needs, reads that section only.

---

Benchmarks

Verified in live use with Claude Opus 5.5 and Claude Sonnet. Any MCP-capable model can use the tools.

Don't take our word for it. Test the tools free in the hosted sandbox with your own model and your own repository, then post your results in Discussions → Benchmarks: model, task, tokens, time.

Found a bug? Open an issue with the steps to reproduce it. Every report makes the tools better.

---

Connecting

The sandbox server speaks standard MCP over HTTP POST with SSE support.

Compatible with any MCP client. Used in practice through Claude.ai and through direct API integration (our benchmark harness).

Hosted sandbox (free live alpha): add https://mcp.itamos-technologia.com/mcp as a connector in any MCP client. A page opens where you create your sandbox with one click, no account needed. Sandboxes are deleted after 10 minutes of inactivity, so don't use them for sensitive data.

---

Self-hosting

Requirements

• Linux with ZFS (required). Every sandbox is its own ZFS dataset with a hard size limit (quota) and compression, so no user can fill the disk for everyone else. The server checks this at startup and refuses to start without it.
• Node.js 20 or newer, plus build tools for the native modules (Debian/Ubuntu: apt install build-essential python3).
• git for the git tool, and Chrome or Chromium for web_skeleton.

Install

git clone https://github.com/itamos-technologia/itamos-mcp-tools.git
cd itamos-mcp-tools
npm install

One npm install sets up everything, including the tools folder.

On Ubuntu 26.04 you can install the package from Releases instead: sudo apt install ./itamos-mcp-tools_1.1.0_all.deb. It sets up a service user, a systemd service, settings in /etc/itamos-mcp-tools/env and the command itamos-mcp-create-slots, then tells you the next two steps.

Create the sandbox slots

sudo scripts/create-slots.sh tank/sandboxes 500 2G $USER

This creates slot_001 to slot_500 under the ZFS dataset tank/sandboxes (use your own pool name), each with a 2 GB quota and lz4 compression, owned by the user the server runs as. It is safe to run again.

Start

SANDBOX_ROOT=/tank/sandboxes SANDBOX_TOTAL_SLOTS=500 npm start

From the same machine, add http://localhost:4200/mcp as a connector in your MCP client. Local connections are identified by IP address and need no sign-in.

Going public

Firewall port 4200 and put an HTTPS reverse proxy (for example nginx) in front of it that sets X-Forwarded-For and X-Forwarded-Proto. Requests that arrive through the proxy use the anonymous one-click sign-in, so users on shared addresses (such as claude.ai) each get their own sandbox. Direct connections to port 4200 skip the sign-in, which is why the port must not be reachable from outside.

Text models (optional)

read_file and web_skeleton give untitled paragraphs and page sections a short title (5 to 10 words) from a small summarizer model, so the agent can tell parts apart without reading them. Without a summarizer everything still works; those parts are then named by their first line.

Any OpenAI-compatible chat endpoint works, set with ITAMOS_SUMMARIZER_URL (default http://127.0.0.1:8090): a single llama-server, Ollama, or the included router in front of several GPUs. The whole paragraph is always sent, never cut.

The hosted sandbox runs Gemma 4 E4B (Q4_0) as summarizer and EmbeddingGemma 300M as embedder on three GPUs (one 16 GB MI50 and two 8 GB V340 dies), each summarizer started like this:

llama-server -m gemma-4-E4B-it-Q4_0.gguf -ngl 99 --port 8190 \
  -np 64 -c 524288 --kv-unified -fa on --cache-type-k q4_0 --cache-type-v q4_0 -rea off

• --kv-unified: one shared KV pool, so each request uses only the tokens it needs.
• --cache-type-k q4_0 --cache-type-v q4_0 (needs -fa on): a 4-bit KV cache, about 4.5 KB per token for this model.
• -c is the pool in tokens: 524,288 on the MI50, 262,144 on each V340 die (with 32 slots).

The router (router/llm-router.js, service itamos-llm-router) listens on 8090 (summarizer) and 8091 (embedder). Before sending a request it counts its tokens with the backend's own tokenizer, sends it to a GPU whose pool has room (preferring faster GPUs by weight), and makes it wait in line when none has. A GPU that fails is skipped for 15 seconds and the request is retried on another. Configure it with ROUTES (in the package: /etc/itamos-mcp-tools/router.env); GET /router/status shows each GPU's pool, reserved tokens and queue.

The embedder on 8091 serves other Itamos services; the tools don't use embeddings yet.

Settings

Variable	Default	What it does
SANDBOX_ROOT	/fast/sandboxes	Folder holding the slot datasets
SANDBOX_TOTAL_SLOTS	500	Number of slots
SANDBOX_PORT	4200	Port the server listens on
SANDBOX_TTL_MINUTES	10	Idle minutes before a sandbox is emptied
SANDBOX_DATA_DIR	./data	Where sign-in data is kept (hashed)
PUBLIC_BASE_URL, PUBLIC_MCP_URL	from the proxy headers	Public addresses, if the proxy can't supply them
WEB_SKELETON_PUBLIC_ONLY	off	Set to 1 on a public server: web_skeleton then refuses private and local addresses
ITAMOS_SUMMARIZER_URL	http://127.0.0.1:8090	Summarizer for titles (any OpenAI-compatible chat endpoint). Empty turns titling off
ITAMOS_SUMMARIZER_MODEL	summarizer	Model name sent with each request; only Ollama needs a real one
SANDBOX_SMTP_CREDENTIALS	a path on the Itamos server	JSON file (email, app_password, smtp_server, smtp_port) for the optional "your sandbox is ready" email. Without it the email is not sent and the server logs why.
ITAMOS_COSTS_DB	a path on the Itamos server	SQLite file for token-savings statistics. If it can't be written, statistics are skipped.

---

License

Dual licensed. Free under AGPL-3.0: use, modify and share, with source published for modified versions, including network use. Building it into a closed product? A commercial licence removes the AGPL obligations. Contact info@itamos-technologia.com.

---

About

Designed and built by Konstantinos Karamperis, founder of Itamos Technologia in Trikala, Greece. A systems architect with a background in industrial engineering and infrastructure, building AI-native developer tools on AMD hardware with open-source inference stacks.

LinkedIn · itamos-technologia.com