Sema: When the Hash Is the Word

Content-addressed semantics for multi-agent coordination.

Sema is a semantic commons that content-addresses meaning itself: the definition is the identifier. By deriving identifiers from the cryptographic hash of a pattern's definition, any divergence in meaning produces a distinct hash, guaranteeing that misaligned agents halt rather than fail silently.

Web: semahash.org

Install

Claude Code (zero-install via uvx)

The fastest way to add sema to Claude Code — no prior pip install needed:

claude mcp add sema -- uvx --from semahash sema mcp

This uses uv to download, install, and run sema in an isolated environment on first invocation, then caches it for subsequent calls. Same shape as npx -y in the Node ecosystem.

Permanent install (pip)

pip install semahash

For MCP server support (recommended for AI agents):

pip install "semahash[mcp]"

Then register with Claude Code:

claude mcp add sema -- sema mcp

Quick Start

Use with AI Agents (MCP)

Already covered above via the uvx or pip install paths. To add Sema to OpenClaw via mcporter:

npm install -g mcporter
mcporter config add sema --command sema --arg mcp --scope home

Or clone and install locally for development:

git clone https://github.com/emergent-wisdom/sema.git
pip install -e "./sema[mcp]"
claude mcp add sema -- sema mcp

Your agent now has access to sema_search, sema_lookup, sema_handshake, and 5 more tools. Any MCP-compatible framework works — Sema exposes a standard stdio server.

Verify it works — ask your agent: "Search sema for coordination patterns and handshake on StateLock"

Full integration guides: Claude Code | OpenClaw

Use via CLI

# Search the vocabulary
sema search "coordination"

# Look up a specific pattern
sema resolve StateLock

# Print a pattern's full definition
sema show StateLock

# Browse the graph structure
sema skeleton

# Start local API + web frontend (binds to 127.0.0.1 by default)
sema serve

Bring Your Own Vocabulary

Build a private registry from scratch — no PR or maintainer in the loop:

sema init ./mylib.db
export SEMA_DB_PATH=$(pwd)/mylib.db
sema apply --add path/to/MyPattern.json
sema search "..."

Subsequent sema commands (including sema mcp) read from your private registry. See CONTRIBUTING.md for the canonical contribution path and docs/versioning.md for the refinement and supersession policy.

Use in Python

from sema.core.actions import sema_handshake
import json

# Look up the canonical hash
result = json.loads(sema_handshake("StateLock"))
print(result["canonical_ref"])  # StateLock#7859

# Verify alignment
result = json.loads(sema_handshake("StateLock#7859"))
print(result["verdict"])  # PROCEED

Try the Protocol (No API Keys Needed)

python experiments/demos/local_handshake.py

See the handshake in action: matching hashes PROCEED, mismatched hashes HALT, unknown patterns HALT. Takes 2 seconds.

How It Works

word = hash(canonical(definition))

Take any concept (a coordination protocol, a reasoning pattern, a trust mechanism), express it in canonical form, hash it. That hash IS the word. Change one byte in the definition, get a different word.

Agent A: "Let's use StateLock#7859"
Agent B: sema_handshake("StateLock#7859")
         -> PROCEED (hashes match) or HALT (drift detected)

This is the Anti-Postel principle: same bytes = PROCEED, different bytes = HALT. No ambiguity, no silent failures.

The Vocabulary

453 patterns across 13 categories and 4 layers:

• Physics — Immutable substrate (locks, entropy, causality)
• Mind — Hybrid cognition (reasoning, inference, strategy)
• Society — Multi-agent coordination (economics, governance, protocols)
• Infrastructure — Operational constraints (data structures, verification)

Each pattern is an executable specification containing machine-verifiable contracts, invariants, failure modes, and typed dependencies.

MCP Tools

When running as an MCP server (sema mcp), these tools are available:

Web Frontend

pip install "semahash[api]"
sema serve
# Open http://localhost:3000

Interactive 3D graph visualization, pattern browser, and search. Built with React + Three.js.

Experiments

The experiments/ directory contains a controlled multi-agent design challenge comparing three conditions:

Agents with Sema patterns produced physics-grounded designs that survived adversarial scrutiny. Agents without Sema produced shallow designs that failed safety review.

To reproduce:

cd experiments/sema_design_challenge
export GOOGLE_API_KEY=your_key
./reproduce.sh

See experiments/sema_design_challenge/README.md for details.

Key Properties

• Zero semantic collisions across the full vocabulary
• 16.9x average token compression via content-addressed stubs
• Fail-closed architecture — mismatches halt, never fail silently
• Mean embedding similarity of 0.21 — high structural distinctness

Using with understanding-graph

Sema gives your agents shared semantic memory — a vocabulary of cognitive patterns with content-addressed identity. Understanding Graph gives them shared episodic memory — the actual thinking trail behind a decision. They compose:

claude mcp add sema -- uvx --from semahash sema mcp
claude mcp add ug   -- npx -y understanding-graph mcp

With both installed, an agent can:

1. Anchor an understanding-graph decision node in a sema pattern hash (e.g. StateLock#7859) so the meaning of the primitive can never drift.
2. Use graph_semantic_search to find all past graph nodes that reference a given sema pattern — hash-stable history, not keyword matching.
3. Call sema_handshake before writing a decision that depends on a shared concept; if it returns HALT, the agent writes a tension node instead and stops, preventing silent divergence.

Full walkthrough: docs/using-with-understanding-graph.md

Repository Structure

sema/
├── src/sema/              Core library (hashing, validation, MCP server, API)
├── data/                  Vocabulary (453 pattern cards + taxonomy database)
├── docs/                  Documentation (philosophy, schema spec, CLI reference)
├── paper/                 Academic paper (sema.tex)
├── web/                   Web frontend (React + Three.js graph visualization)
├── experiments/
│   ├── emergent-swarm/    Multi-agent engine (bundled for experiment reproduction)
│   ├── sema_design_challenge/  Main experiment (3 conditions, 5 runs, full traces)
│   └── demos/             Standalone demos (local handshake, Babel Test)
├── integrations/          Integration guides (Claude Code, OpenClaw, any MCP client)
└── pyproject.toml         Package config (extras: [mcp], [api], [full])

Contributing

Want to add patterns, improve existing ones, or host the frontend locally? See CONTRIBUTING.md.

Citing

@unpublished{westerberg2026sema,
  title={Sema: Content-Addressing Meaning for Safe Multi-Agent Coordination},
  author={Westerberg, Henrik},
  year={2026},
  note={Preprint}
}

License

Sema is dual-licensed:

• Code (everything in src/, web/, experiments/, scripts/, and the package config) — MIT. Self-host it, fork it, build commercial products on top of it.
• Content (the pattern vocabulary in data/, the documentation in docs/, the academic paper in paper/, and the prose displayed on semahash.org) — CC BY 4.0. Reuse the patterns and prose anywhere, for any purpose including commercial, as long as you attribute Henrik Westerberg.

For academic citation, see CITATION.cff. GitHub renders this as a "Cite this repository" button on the project page that generates APA and BibTeX automatically.