assemble_context: a ready-to-read context pack

Behaviour

Returns the corroborated, story-deduped, trust-ranked passages for a question, packed to a token budget, each carrying the reason it was included.

The engine runs the whole retrieval loop — search, dedup, corroborate, rank, pack — so you just read. Use this when the question is 'what should I know about X' rather than 'find me the page that says Y'.

`budget` is in tokens and is honoured as a ceiling; `estimated_tokens` reports what was actually packed.

Costs 3 credits; a call that runs and fails costs 1.

Arguments

inputSchema

Name	Type	Description
`budget`	integer	Token budget for the pack
`mode`	"keyword" \| "semantic" \| "hybrid"	—
`q`required	string	The question

Returns

A JSON payload in the text content block, and the same payload as structuredContent for clients that support it.

outputSchema

Name	Type
`estimated_tokens`required	integer
`items`required	object[]
`query`required	string
`token_budget`required	integer

Calling it

Over Streamable HTTP, with the required arguments filled in:

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "assemble_context",
    "arguments": {
      "q": "<q>"
    }
  }
}

Most clients do this for you — see Connect a client. To drive it by hand, complete the startup lifecycle first.

The same call over REST

If your client does not speak MCP, this capability is GET /assemble_context, taking the same parameters and returning the same body.

curl -H "x-api-key: $UNLOB_API_KEY" \
  "https://api.unlob.com/assemble_context?q=<q>"