Memory that keeps its word: memory, obligations and proofs in one MCP endpoint

A developer released soma-protocol, an MIT-licensed MCP endpoint that combines agent memory, signed obligations and offline-verifiable proofs in one service. The system stores facts outside the context window with BM25, stemming and tag search merged via reciprocal rank fusion, creates Ed25519-signed envelopes for tasks with token budgets and mandates, and lets anyone verify signatures offline since the public key travels with the signature. A ten-question bilingual recall test reports recall@3 of 10/10, and the project credits Hindsight (MIT) for its multi-channel retrieval and observation ideas.

Most agent memory systems answer one question: what did we say before? That is useful, and it is also where most of them stop. An agent that remembers a fact but cannot be held to a promise is still an agent you have to babysit. So we built the boring part too. Memory. soma remember stores a fact with tags and importance. soma recall searches by meaning — full-text BM25, word stems and tags merged with reciprocal rank fusion. soma reflect builds an observation and records which facts it came from derived from edges . Facts live outside the context window, so long memory does not mean a long prompt. Obligations. soma send creates a signed envelope: goal, task, recipient, token budget, mandate, price — all named before work starts. soma claim takes it exclusively; soma complete closes it with a signed receipt. A refusal always names the rule that rejected it. Proofs. Every envelope is signed with Ed25519. soma verify checks the signature offline, because the public key travels with the signature. No call back to the node, no trust in our server required. A memory without a benchmark is a belief. We keep a small honest one: ten bilingual questions Russian and English , and the expected answer must be in the top 3. tests/test recall quality.py — recall@3: 10/10 That test caught two real failures: Russian word forms "подпись" vs "подписан" and English cost questions ranking a tool description above the price. Both are fixed in the ranking, not in the test. The ideas for multi-channel retrieval, temporal filters and observations came from reading Hindsight MIT . Credit where it is due: the category moves fast, and the honest thing is to measure against it. soma export returns every fact and every envelope as one JSON. No lock-in, no "we will delete it for you". If the node is not useful, you take your memory and go. Source: https://github.com/mrpkk/soma-protocol https://github.com/mrpkk/soma-protocol MIT . MCP registry: io.github.mrpkk/soma-protocol . If you run agents that must remember, promise and prove — the demo is one click away, and I would rather hear what is missing than what is nice.