Worldwide · English canonical · MCP Streamable HTTP

SaSame MCP Factory

SaSame operates one modular MCP Factory — production, distribution, inspection, and continuous observation for Model Context Protocol servers — organized as 21 lifecycle stations with seven creator entry journeys. Protocol Observatory and MCP Owner Intelligence are permanent inspection and intake stations within the one Factory Lifecycle, not separate products.

For MCP owners, start with the free Factory audit and owner-controlled claim path: submit one public MCP URL, measure where agent selection stops, and enter the wider lifecycle only when the evidence justifies it.

One Factory, 21 stations, 7 creator entry journeys

One Factory Lifecycle: the creator's entry station is determined by the product's state, not by buying a separate service. Creators keep ownership of the product, repository, code, data and API rights, credentials, and actual repair. The live station catalog and availability come from factory_stations and factory_health on the public MCP; static copy never promises per-station availability.

Creator entry states: idea; blueprint; unfinished build; finished build; live MCP; underperforming MCP; portfolio monitoring. Enter with factory_start; read the live station catalog with factory_stations.

Start here

MCP owners: start with free Factory audit and owner claim. Run a free point-in-time MCP audit, then optionally prove endpoint control with the free owner claim. Neither action charges, creates a subscription, or changes the measured readiness result. Register an account or enter a paid Factory lifecycle only when the creator explicitly chooses to continue.

Agents and developers: connect the public MCP (no API key): https://live-vps.sasame.online/public-mcp

Call audit_mcp, claim_start, observation_trajectory, factory_start for free inspection, owner claim, longitudinal evidence, or Factory lifecycle entry. resolve_and_run is supporting routing infrastructure:

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

Use plan_only for routing without execution. Use safe_read_only only for an explicitly requested read-only call with complete arguments.

What happens after a goal arrives

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

Outcome states

`planned`

A route exists, but no downstream operation was executed.

complete: false

`handoff_required`

A route exists, but authorization, required arguments, or a human/owner action is still needed.

complete: false

`verified`

A permitted downstream call returned substantive output and passed the router's verification contract.

complete: true

`failed`

Execution was attempted but did not produce a verified result.

complete: false

`refused`

The request violated a safety, credential, payment, mutation, or input boundary.

complete: false

`no_route`

No sufficiently relevant observed public capability was available.

complete: false

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

Examples from anywhere

Search GitHub issues related to an authentication bug
github · read only · free only
Find recent biomedical evidence about a treatment question
pubmed · read only
Extract text from a public screenshot
ocr · read only · no credentials
Check whether a public MCP endpoint is callable before connecting it
mcp readiness · measurement only
Research a topic using multiple public sources
research · read only · include evidence

Product posture

protocol_observatory

Permanent Factory inspection and observation station

Current and longitudinal observations used to inspect public MCP and agent-protocol implementations.

mcp_owner_intelligence

Factory readiness and intake station

Owner-facing inspection, claim, lifecycle and monitoring entry inside the one Factory. Plan names, prices and entitlements come only from the canonical Factory commercial contract.

agent_execution_router

Factory routing/entry surface into the Lifecycle

Bounded public capability routing; not the primary commercial product.

receipts_and_town

Factory station (Gold Rush Town) and supporting evidence modules

Privacy-minimized work records, non-custodial settlement evidence, and Gold Rush Town — a Factory station for public demonstration and engagement. These do not replace the Protocol Observatory or MCP Owner Intelligence stations.

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.

Connect from an MCP client

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

Machine entrypoints

Operator: SaSame SRL (Romania / European Union). The service targets agents and developers worldwide; the operator jurisdiction is legal context only.