Signal Canvas
Signal Canvas is the citation-first research cockpit in ScienceToStartup. It supports public corpus and paper mode, authenticated workspace mode, export, save-to-workspace, and deep links into Build Loop and Talent.
Public. Best for broad research synthesis, opportunity ranking, and next-action generation.
Public. Anchors the answer to the canonical paper kernel and evidence receipt before synthesis.
Authenticated. Reuses saved thesis context, prior evidence queries, linked papers, and thread history.
source_context = {
surface,
mode,
query,
normalized_query,
paper_ref,
thread_id,
query_id,
workspace_id
}The same object powers Build Loop, Talent, workspace save/reopen, browser deep links, capability discovery, and agent documentation.
Corpus mode
curl -X POST https://sciencetostartup.com/api/research/search \
-H "Content-Type: application/json" \
-d '{"query_text":"What are today\'s highest viability papers?","time_scope":"weekly","match_count":10,"mode":"corpus","allow_clarification":true}'Paper mode
curl -X POST https://sciencetostartup.com/api/research/search \
-H "Content-Type: application/json" \
-d '{"query_text":"What should I verify before building this?","time_scope":"weekly","match_count":10,"mode":"paper","paper_ref":"2604.01224v1","allow_clarification":true}'{
"jsonrpc": "2.0",
"id": 7,
"method": "tools/call",
"params": {
"name": "search_signal_canvas",
"arguments": {
"query": "What are today's highest viability papers?",
"scope": "weekly"
}
}
}