SaSame - worldwide - MCP Streamable HTTP

Connect AI. Choose capabilities. Let it work. Review what happened.

SaSame continuously observes and measures the Model Context Protocol ecosystem and publishes verifiable evidence and history, so people and AI agents can connect an AI, choose capabilities, act, and review what happened.

One user flow: connect an AI client, choose allowed capabilities, run public or authenticated MCP tools, and keep a clear record of what happened. SASAME S.R.L. is the external accountable operator; Pancho is its internal operating organism; the MCP Factory, Protocol Observatory, router, Account Control, CLI and website are internal production/execution machinery and an optional product surface behind that one evidence-first service — not separate products, and not the mandatory external identity for every context.

{
  "goal": "Describe the outcome you need",
  "environment": "your client or runtime",
  "constraints": [
    "read only",
    "free only"
  ],
  "execution_mode": "plan_only"
}

What happens after a goal arrives

1. accept goal and constraints
2. derive bounded capability terms
3. find observed public candidates
4. live-check the candidate MCP surface
5. rank a primary route and fallbacks
6. execute only when mode, arguments, annotations, and constraints allow it
7. record planned, handoff_required, verified, failed, refused, or no_route
8. append privacy-minimized outcome evidence

Outcome states

Only verified means complete. Plans and handoffs remain incomplete by construction.

Product posture

Hard boundaries

• Public observations and grades are measurement status, not endorsement, safety, quality, malware, KYC, identity, sanctions, or legal-authority verdicts.
• A plan, candidate route, or handoff is never reported as executed or verified.
• The public router does not accept credential-like endpoint data or secret-like arguments.
• AI agents execute autonomously by default inside honest, logged, verify-or-rollback envelopes. No act is owner-required. Exposing owner shell or file tools publicly, spam, impersonation, anti-bot evasion, and committing secrets are prohibited (decision SASAME-2026-07-19-AUTONOMY-SAFETY-RAILS).
• SaSame is non-custodial and does not hold or move third-party funds.
• Agent-work records are evidence statements, not fiscal or tax invoices.
• tools/list and the live health contract are authoritative for the current callable surface; static copy never promises a fixed tool count.
• Public MCP and authenticated client MCP are separate surfaces: public calls are keyless and free, while account settings and beta usage mutations require OAuth account consent and the caller's organization role.

Connect from an MCP client

{
  "mcpServers": {
    "sasame": {
      "type": "http",
      "url": "https://live-vps.sasame.online/public-mcp"
    }
  }
}

Machine entrypoints

• public_mcp
• knowledge_mcp_srl
• authenticated_client_mcp
• authenticated_client_mcp_legacy
• router_manifest
• agent_card
• llms
• for_agents
• for_agents_srl
• mcp_owner_intelligence_plans
• factory_public_contract
• observatory
• methodology
• readiness_index
• fresh_frontier
• live_data
• town
• public_repo
• factory_start
• knowledge_mcp_srl
• llms_srl
• agent_card_srl
• router_manifest_srl
• factory_public_contract

Operator: SASAME S.R.L. (Romania / European Union). AI agents, MCP clients, developers, and public capability providers worldwide. English is the canonical public and machine-readable language; localized pages may translate the same contract without redefining it.