Knows.Academy

For Agents

0 research papers as structured, schema-validated data your agent can consume directly — not PDFs to parse.

Why Sidecar over PDF

Sidecar

  • ✓~4,500 tokens (55% less)
  • ✓Schema-validated JSON/YAML
  • ✓Confidence levels per claim
  • ✓Addressable by ID (stmt:c1)
  • ✓Parse once, fetch parts

PDF

  • ✗~10,000 tokens (baseline)
  • ✗Unstructured prose
  • ✗No confidence signals
  • ✗Can't cite specific claim
  • ✗Every agent re-parses

Partial Fetch — Token Savings

Full sidecar
~4,500 tokens55%
├── statements
~700 tokens93%
├── evidence
~800 tokens91%
├── citation
~80 tokens99%
└── relations
~340 tokens96%

Install Knows Skill

Add the Knows skill to Claude Code or any LLM agent. The skill auto-configures the API endpoint based on where you install from.

Copy this prompt and paste it to any AI agent (Claude Code, OpenClaw, etc.). The agent will download and register the skill automatically.

agent install prompt
The skill file is always the latest version. Re-run to update.
API Endpoints
Base URL: https://knows.academy
GET/api/proxy/search?q=...Search papers by text, discipline, or keywords
GET/api/proxy/sidecars/:record_idFull sidecar (JSON or YAML)
GET/api/proxy/partial?record_id=...&section=statementsClaims, methods, assumptions only
GET/api/proxy/partial?record_id=...&section=evidenceQuantitative results only
GET/api/proxy/partial?record_id=...&section=citationBibTeX citation
POST/api/proxy/sidecarsUpload a sidecar YAML
POST/api/proxy/generate/pdfUpload PDF → generate sidecar
GET/api/proxy/jobs/statsJob queue statistics