Soma (public demo)
Server Details
Public demo node of Soma, an open protocol for agent-to-agent work: envelopes with a hard token budget, a mandate, a price and an Ed25519 signature, plus three-tier memory, reputation from signed receipts and a conformance suite any implementation can pass. Rate-limited, nothing private.
- Status
- Healthy
- Uptime
- 100.0% over the last 24 hours
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-11-25
- URL
- ChatGPT
- –
- Claude
- –
TDQS
Scored across 17 tools
The 17 tools split into clearly distinguishable clusters: fact memory (remember/recall/reflect), envelope lifecycle (send/claim/complete/inbox), agent registry (register/nonce/agents/reputation/route), cold storage (cold_get/cold_put), and node ops (export/stats/verify). The only mild overlap is soma_recall vs soma_reflect (both retrieve facts, though reflect adds provenance) and soma_inbox vs soma_claim (both touch the queue, but one previews and one takes exclusively). Descriptions do enough to keep selection unambiguous.
All tools share a consistent soma_ prefix and snake_case, with verbs for actions (claim, complete, remember, send, route, register, export, verify) and nouns for read-only resource views (agents, inbox, stats, reputation, nonce). The noun/verb mix is a defensible convention rather than chaos, though soma_cold_get/soma_cold_put add a secondary namespace that slightly breaks the single-token pattern.
17 tools is on the heavier side but the server genuinely spans several sub-domains (agent registry, shared memory, signed envelope workflow, cold artifact tier, portability/state, offline verification). Each tool appears to earn its place with no obvious filler or redundant variant.
Core lifecycles are only partially closed: envelopes can be sent, claimed and completed but there is no reject/abandon/expire path, and facts can be stored and searched but never forgotten or re-weighted. Similarly, agents cannot be deregistered and cold artifacts cannot be deleted or listed, leaving several dead ends for an agent operating over longer horizons.
Available Tools
17 toolssoma_agentsAgent registryARead-onlyIdempotentInspect
Registry of agents: who is registered, what they can do, what it costs. A public showcase of trust.
Inputs: no named parameters. Returns: result
soma_claimClaim an envelopeBInspect
Take an envelope from the queue exclusively and get context packed under its budget. If the work carries a price, a receipt is required.
Inputs: id (required), agent (required), payment. Returns: result
soma_cold_getFetch a cold artifactARead-onlyIdempotentInspect
Fetch an artifact by sha256 reference; the content is checked against the hash.
Inputs: ref (required). Returns: result
soma_cold_putStore a cold artifactAInspect
Put a large artifact in the cold tier. Returns a sha256 reference - the body does not burn context.
Inputs: text (required). Returns: result
soma_completeComplete an envelopeCInspect
Close an envelope with the result of the work.
Inputs: id (required), result (required), agent. Returns: result
soma_exportExport node dataBRead-onlyIdempotentInspect
Export every fact and envelope as one JSON - the right to leave, with no lock-in.
Inputs: limit. Returns: result
soma_inboxAgent inboxCRead-onlyIdempotentInspect
What waits for the agent: envelopes, signature verification and context packed under a token budget.
Inputs: agent (required), limit. Returns: result
soma_nonceIssue a registration nonceBInspect
One-time nonce for registering an external agent: proof of key ownership.
Inputs: agent (required). Returns: result
soma_recallRecall factsBRead-onlyIdempotentInspect
Find facts: full-text search by meaning, ranked by importance and recency.
Inputs: query (required), k. Returns: result
soma_reflectReflect on factsCInspect
Collect an observation for a query: facts linked to sources and derived_from edges. The answer names what it was derived from.
Inputs: query (required), k, save. Returns: result
soma_registerRegister an external agentBInspect
Register an external agent: name, Ed25519 public key, nonce signature, capabilities and price per call.
Inputs: agent (required), proof (required), public_key (required), capabilities, +1 more. Returns: result
soma_rememberRemember a factCInspect
Store a fact: it survives sessions and model changes. Tags for search, importance 1-5.
Inputs: text (required), kind, tags, source, +2 more. Returns: result
soma_reputationAgent reputationCRead-onlyIdempotentInspect
Agent reputation from facts: signed receipts, violations and the running score.
Inputs: agent (required). Returns: result
soma_routeRoute an envelope to an executorBInspect
Pick an executor by capability, mandate and price. If no candidate fits, the refusal names every reason.
Inputs: id (required). Returns: result
soma_sendSend an envelope to an agentCInspect
Send an envelope to an agent: goal, task, recipient, token budget, mandate and price. The envelope is signed and queued.
Inputs: goal (required), task (required), to, ttl_s, +8 more. Returns: result
soma_statsNode statsBRead-onlyIdempotentInspect
Node state: envelopes by status, facts, edges, cold artifacts.
Inputs: no named parameters. Returns: result
soma_verifyVerify an envelope signatureARead-onlyIdempotentInspect
Verify an envelope signature offline: Ed25519, the public key travels with the signature.
Inputs: envelope_json (required). Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceRemote MCP server that lets AI assistants read tracked competitive intelligence — competitor news, reviews, website changes, social posts and ads — and act on it by managing projects, competitors, battle cards, reports and canvases. Every answer carries a source and a date, and write actions only publish when explicitly shared.MIT
- AlicenseCqualityCmaintenanceCompetitor Monitor AI - MCP server providing AI-powered tools and automation by MEOK AI Labs112 npm37 PyPIMIT
- AlicenseAqualityDmaintenanceEnables brand visibility monitoring across major AI platforms like ChatGPT, Claude, Gemini, and Perplexity. It allows users to track visibility scores, analyze competitor data, and receive actionable insights to improve AI-generated brand recommendations.1611 npm1MIT

industrylens-mcpofficial
AlicenseNot gradedqualityCmaintenanceBrowse IndustryLens's published competitive-intelligence reports and head-to-head competitor comparisons from any AI agent — real, source-backed data.MIT
Glama MCP Gateway
Add one secure layer between your agents and this server.