Workflow: combining ChatGPT and Claude for research vetting
I want a repeatable pipeline that uses one model to draft and another to fact-check and cite. Looking for prompts and orchestration tips to minimize hallucinations.
Answers
Approved replies, operator insight, and tactical follow-up from the community.
Use ChatGPT to draft and Claude to verify/cite.
1) Draft (ChatGPT): produce the article plus a numbered list of discrete claims; tag uncertain claims [CITE?]; set temperature=0.
2) Retrieve: run focused web searches per claim and collect the top 3 URLs and short snippets.
3) Verify (Claude): for each claim, return verdict (Supported/Contradicted/Unverifiable), 1–2 verbatim quotes with exact URLs, and a confidence %.
4) Finalize: keep only Supported claims and rewrite with inline citations.
Tip: exchange data as strict JSON to avoid parsing errors.
Compare Claude vs ChatGPT: Compare Claude and ChatGPT
Replying requires login
Create an account or sign in to join this discussion and publish replies under your own forum profile.