How to Migrate ChatGPT Conversations to Claude or Gemini

You’ve invested two months in a ChatGPT conversation working through a complex problem. Now you want to switch to Claude for superior long-context handling or Gemini for tighter Workspace integration, without throwing away all that progress. How do you make the jump?

Here are three migration approaches, ranked by fidelity.

1. Full transcript paste (simplest and most effective)

Export your ChatGPT conversation as Markdown. Open a new Claude or Gemini chat. Paste the entire Markdown transcript into a single message with this prompt:

Here's a conversation I had with ChatGPT that I want to continue with you.
Read it, then continue from where we left off.

[paste the full Markdown export]

Continue from this point.

What works:

• Complete reasoning chain preserved
• All code, math, and citations retained intact
• The new model receives the full context ChatGPT had

What doesn’t:

• Extremely long conversations may hit Claude’s 200K or Gemini’s 1M token limit (rare, but possible for multi-month research threads)
• The new model treats pasted history as observed output, not its own reasoning, it may question earlier conclusions
• Initial streaming may slow down while processing large context

This method works best for most use cases. ChatExport AI’s Markdown exports are token-efficient, a typical 50-message conversation runs about ~10K tokens, well within any modern model’s context window.

2. Summary primer (faster but lossy)

When the full transcript exceeds reasonable limits, ask ChatGPT to create a handoff summary first:

Summarise the key conclusions, decisions, and open questions from
this conversation in a way that another AI assistant could pick up
where we left off. Include any code or data that's still relevant.

Take that summary and paste it as the opening message in your new Claude or Gemini chat:

I had a long conversation with ChatGPT that produced this summary.
Here's where we got to. Continue from this point.

[paste summary]

You’ll lose some fidelity, ChatGPT decides what matters. But for strategic discussions and ongoing projects, a 2K-token summary usually captures enough context to continue effectively.

3. JSON export → structured prompt (maximum fidelity)

For research where conversation structure matters: export ChatGPT as JSON. The JSON includes per-turn metadata (role, content, model, timestamp). Filter to relevant turns and build a clean, structured primer:

Here are the turns from a previous conversation that are relevant
to where we're going next:

[paste filtered JSON, manually edited]

Continue from this point.

This approach shines for engagements with 200+ messages where only ~20 turns carry forward, you get surgical precision without noise.

Per-platform migration specifics

Migrating to Claude

Claude Sonnet 5 handles long pasted contexts exceptionally well. It may occasionally ask “Should I treat this as input I’m reading, or as our shared history?”, clarify this in your primer message.

For ongoing work, upload the Markdown export as a Project file instead of pasting. Claude can then reference it across multiple conversation threads.

Migrating to Gemini

Gemini 3 Pro offers a 1M-token context window, accommodating even massive conversation histories. The system instructions field is ideal for summary primers (“background context:

”). With Gemini Advanced, you can also attach the Markdown export as a file.

Migrating to DeepSeek (R1)

Switching to capture reasoning improvements? DeepSeek-R1 will iterate on previous reasoning chains. Be explicit in your prompt: “Here’s where we got to. Re-examine the conclusions critically.” R1 excels at second-pass reasoning that catches what GPT-4 missed.

Migrating to Perplexity

Perplexity is search-grounded, pasting long conversation primers doesn’t align with its UX design. Better approach: extract a single focused research question from your ChatGPT conversation and ask Perplexity fresh. Use ChatGPT’s reasoning to formulate the question; let Perplexity surface current sources.

Common migration mistakes to avoid

Don’t: Screenshot the ChatGPT conversation and upload images to the new chat. The model must OCR your screenshots, which destroys formatting, math notation, and code structure. Markdown export is always superior.

Don’t: Ask the new model to “remember” the old conversation as if it has persistent memory. It doesn’t. Every chat starts fresh with only what you include in the prompt.

Don’t: Expect the new model to reach identical conclusions. Getting a different perspective is often exactly why you’re switching models.

Real-world migration example

I spent a month with ChatGPT designing a Postgres schema for multi-tenant SaaS. I switched to Claude for implementation because Claude’s extended context held the entire design in working memory more effectively.

Migration workflow:

1. Exported ChatGPT conversation to Markdown via ChatExport AI, approximately 25K tokens of design discussion
2. Pasted the full Markdown as the first message in a new Claude Project
3. Prompted Claude: “Read this. List the design decisions you’d push back on.” Received 4 substantive critiques immediately
4. Iterated on those critiques over two weeks. Schema was implemented and tested in Claude with complete design context

That’s effective cross-model workflow, letting each AI do its best work.

Related guides

• ChatGPT vs Claude vs Gemini, when to choose each model
• How to export ChatGPT to Markdown
• Claude export guide