Context
You want to capture context, logic, and setup from another AI tool (e.g. ChatGPT) and load it into Hatz so it can continue working without starting from scratch.
Practical Solution In Hatz
There isn’t a direct 1:1 way to import an AI agent from one platform to another.
However, the following process provides a reliable way to transfer context and functionality.
Step 1: Extract the Agent From the Existing AI
In your current AI platform, give the agent this prompt:
"I need to continue this project in another AI platform. Please provide the system prompts, context, assumptions, rules, and any relevant information another AI would need to pick up where we’re leaving off. Put everything into a single system prompt that I can reuse elsewhere."
Step 2: Save the Output
Save the response as a PDF, or
Copy the full text (whichever is easiest)
Step 3: Set Yourself Up in Hatz
Before recreating the agent, partners should:
Complete the Hatz onboarding learning in the Learning Portal, including the Hatz AI Navigator Course
Become familiar with:
How agents and workflows are built in Hatz
Available tools and assistants
Best practices for structuring prompts and logic
This step significantly improves the quality of the recreated agent.
Step 4: Use The Community Agent Called "Import From Other Platforms"
Go to the Workshop > Discover > Search for "Import From Other Platforms
Upload your PDF and say "Import This"
Answer the follow up questions
Use your new Agent
Outcome
This is not a perfect 1:1 migration
But it preserves context, intent, and behavior
Partners benefit from:
Multi-step workflows
Access to multiple LLMs (not just ChatGPT)
Better long-term scalability inside Hatz
In many cases, the result is a more powerful and flexible solution than the original setup.
Importing your ChatGPT memory
If you've built up saved memories in ChatGPT, you can bring them into Hatz so your assistant keeps that context.
In ChatGPT, ask it to export your memories as a Markdown document. A prompt that works well:
"Please create a markdown document with all of my memories. Format each memory as a separate bullet point with the date the memory was saved in brackets after the bullet point, followed by a hyphen and the memory on the same line."
Save the result as a .txt or .md file.
In Hatz, go to your workspace Memory settings and use the import option to upload that file.
Review the memories before saving. Each individual memory can be up to 280 characters; longer entries are flagged during review so you can shorten them first, and duplicates of memories you already have are flagged so you don't import them twice.
Migrating from Claude (Projects, apps, and workflows)
There isn't a direct one-to-one importer for Claude Projects, Claude apps, or other Claude setups, but you can carry over the context and logic with the same approach used above:
In Claude, ask the project or assistant to summarize everything another AI would need to continue the work — its instructions, context, rules, and assumptions — into a single reusable system prompt, and save it as a PDF or copy the text.
For a Claude Project, also gather the project's knowledge files or documents so you can add them as sources on your Hatz agent.
In Hatz, use the "Import From Other Platforms" agent in Workshop > Discover to turn that export into a starting agent, then refine it.
As with ChatGPT, this isn't a perfect one-to-one migration, but it preserves the intent and behavior and lets you rebuild the setup as a Hatz agent or multi-step workflow with access to multiple models.