Compatibility you can trust · for people + AI agents

Know what fits
before you buy it.

Tethered Trust checks whether a part, accessory or appliance actually works with what you have, using model numbers, specs and dimensions, and shows its evidence.

Try:

Answers come from Tethered Trust's compatibility graph and fixed rules. When specs come from your text or from AI, the result says so.

One question. An answer with evidence.

Tethered Trust isn't a product search engine. It answers one high-value question: will this exact thing work with that exact thing?

1

Identify the exact item

Model numbers, suffixes and revisions matter more than product names. Tethered Trust matches them against its graph.

2

Apply the rules

Memory generation and form factor, M.2 length, key and interface, USB-C wattage, clearances. Fixed rules, not guesses.

3

Return a decision

Compatible, not compatible, or not sure yet, with confidence, caveats, sources and what to ask next.

Built for the expensive little mistakes.

Every category has purchases that look right until one tiny detail makes them wrong.

Computers

RAM (DDR3–DDR5, SODIMM/DIMM), M.2 SSDs (2230–2280, NVMe/SATA) and USB-C chargers.

Appliances

Replacement parts such as water filters, plus whether a new appliance fits its opening with real clearances.

Growing every day

Every question Tethered Trust can't answer yet is logged, and the most-asked products are added first.

Agent-ready from day one.

The same engine powers this page, an MCP server and an HTTP API, so AI assistants can verify a fit before they buy anything for someone.

Connect an assistant

MCP server

Add this as a custom connector in Claude, ChatGPT or any MCP client:

/mcp

1

check_compatibility: part vs device, with evidence.

2

find_parts: what fits this model.

3

check_fit: will the appliance fit the space.

4

report_outcome: did it actually fit? It improves the graph.

Free during early access. Per-call pricing will use x402 (HTTP 402) payments. API description · llms.txt · API & Agent Terms

POST /v1/check
{
  "existing": "Lenovo ThinkPad T480",
  "candidate": "Crucial CT16G4SFRA32A"
}

→

{
  "verdict": "compatible_with_caveats",
  "fit": true,
  "confidence": 0.94,
  "verified": true,
  "requirements": {
    "form_factor": "SODIMM (laptop)",
    "memory_type": "DDR4"
  },
  "caveats": ["It will run at the device's 2400 MT/s…"],
  "evidence": [{ "type": "source", "text": "…Lenovo PSREF…" }]
}
The long-term idea

A universal physical compatibility graph.

Products, parts, accessories and spaces linked by verified "works with / doesn't work with" relationships, usable by people, retailers and autonomous agents.

Try it