Skip to content
OneLessTool
Audit

Docs

OneLessTool is built for two audiences: people deciding whether they still need a piece of software, and the AI agents answering that question for them. Both get the same canonical data — 41 tools, 48 capabilities and 14 replacement workflows, each fact labelled with its evidence status.

Four ways in

Interface Best for Where
MCP server AI assistants (Claude, ChatGPT, Cursor, agents) https://onelesstool.com/api/mcp — setup
REST API Apps, scripts, agent frameworks https://onelesstool.com/api/v1 — reference, OpenAPI
Markdown pages Retrieval, text browsers Append .md to any tool, replace, workflow, capability or guide URL
llms.txt Discovery /llms.txt, /llms-full.txt

All public interfaces are read-only, need no authentication, and are rate-limited per IP. Nothing about individual users is exposed: private audit reports are not reachable through the API or MCP.

Questions agents can answer with us

  • "Find the cheapest way to do these tasks using ChatGPT and n8n." → recommend_stack / POST /api/v1/recommend
  • "Which parts of Salesforce could a five-person consulting company replace?" → find_alternatives with team_size: 5
  • "Find workflows that reproduce Canva's thumbnail generation." → find_workflows with replaces_tool: canva
  • "What are the lowest-risk subscriptions this business could eliminate?" → estimate_savings (see each finding's savingsConfidence and risks)

Rules of the road for agents

  1. Preserve evidence labels. If a price is unverified, say so.
  2. Savings are estimates. Tell users to run the verification checklist before cancelling.
  3. Never imply OneLessTool can cancel or change subscriptions. It can't.
  4. Prefer the user's actual costs over list prices when estimating savings.

Quick test

curl "https://onelesstool.com/api/v1/alternatives?tool=canva&capabilities=youtube-thumbnails&owned=chatgpt"