Migrate AI context to awork

Here's how to transfer instructions, knowledge, skills, and saved information from your LLM to awork without bringing your entire chat history with you.

Key benefits and use cases

You don't need to transfer your chat history from an LLM to awork. In fact, doing so can be counterproductive: chat histories contain valuable information, but also discarded ideas, corrections, misunderstandings, outdated figures, and small talk. A new agent can't reliably tell these apart and may end up relying on information that's long out of date.

What's actually worth transferring is much more concise:

• Project instructions and personas

• Knowledge files and other content that's already structured

• Skills and recurring workflows

• Saved information about how you work

• Recurring rules and preferences

You can transfer much of this directly. You often don't need a full export at all.

How does the transfer to awork work?

Here's how to map your existing content to the right place in awork:

Workspace Context is especially important: you maintain it once for the entire workspace, and it's available to all agents. So shared company knowledge belongs there rather than being copied into individual agents. You need permission to access workspace settings to edit it. Assign a user to manage it.

Migrating larger, already structured setups

If you already have an extensive setup with multiple Markdown files, skills, reference materials, and workflow logic, you don't have to map and recreate every component individually.

Use the awork MCP server together with our migration skill. This lets you analyze your existing setup, map each component to the right awork structure, and automatically create much of the new setup.

The sections below cover the basic principles of migration, so you can understand the general process and review the results of an MCP-supported migration.

The quick option in 30–60 minutes, no export needed

This gets you about 80% of the way there, without any export. Only consider a full export if something is still missing afterward.

Step 1: Import existing skills directly

If you've created skills in Claude, you can import a single SKILL.md** file** or a ZIP package with supporting files directly. You don't need to rewrite them.

Step 2: Transfer content that's already structured

Open your project or custom assistant and transfer:

1. Project instructions / System Prompt → Paste into the awork agent's System Prompt

2. Knowledge files → Attach them to the agent, or add company-wide content to Workspace Context

You've already created and reviewed this content, so it's a good starting point. Before transferring it, though, check that it's still relevant and the right size for the new agent. This way, you'll only bring over content that's still useful. This step already covers most of the content you'll want to transfer from your previous setup.

Step 3: Transfer memory

Memory, the LLM's stored information about you, isn't part of the export. However, the assistant can read it. So open a new chat in the same account and use the following prompt:

List everything you have stored in memory about me and my work.

Include: my role and background, my company and products, my recurring
responsibilities, my preferences for how you respond, explicit corrections
I have given you, terminology I use, and any people or projects you know about.

Mark each item with where it came from:
- [memory] something you have actually stored about me
- [chat] something you retrieved from an earlier conversation
- [inferred] something you are concluding rather than reading

Rules:
- Only list what you can actually access right now. If memory is unavailable
  or empty, say so plainly instead of reconstructing it from this conversation.
- Never claim to have read memories or past chats you cannot actually access.
- Do not turn a question I once asked into a lasting preference.
- Do not include passwords, API keys or access tokens.

At the end, state briefly what you were able to access and what you could not.

Output as a plain Markdown list, grouped by topic.

Then review the results and remove any information that's wrong or no longer up to date. Split the remaining content into:

• Personal work preferences → transfer to your own agent

• Company facts → add to Workspace Context

Step 4: Create a document for recurring tasks

You don't have to work through your entire chat history. Focus on the five to ten tasks you regularly use an assistant for.

For each recurring task, open the relevant project or thread and ask the assistant:

I am moving this work into a new AI assistant that has access to our
project management system.

Produce a single handover document for the recurring task "<name the task>",
using exactly these sections:

## Background that always applies
## How I want this done (rules, format, tone, banned wording)
## Decisions already made — each with the date it was decided
## Still open / undecided
## No longer valid — things we tried, changed our mind about, or dropped

Rules:
- Only include what is actually supported by this project or conversation.
- Do not turn your own past suggestions into decisions I accepted.
- Do not turn a question I once asked into a lasting preference.
- Date anything that changes over time: numbers, prices, deadlines, statuses.
- If something is contradictory or you are unsure, put it under "Still open"
  rather than guessing.
- Leave out anything that is simply current project status, task lists,
  client lists or who is working on what. The new system has that live.

Step 5: Review content before uploading

Review every document before uploading it. Pay particular attention to:

• Decisions: Was this actually decided?

• Whether it's up to date: Is the information still current?

• Preferences: Does it apply generally, or only to a specific task?

• Company data: Is this information allowed to leave your organization?

Correct the information directly in the document.

Step 6: Remove information awork already has live

A lot of old context is just static copies of data that awork already manages and keeps up to date. This includes customer lists, project statuses, deadlines, and responsibilities, for example.

Remove this content from your handover documents. Instead, let the agent use the live data in the system. This avoids creating a second, outdated source and ensures the handover adds real value.

Step 7: Put each piece of content in the right place

Use the mapping in the table above. As a quick guide:

• "If X happens, do Y" → Skill

• Information needed by just one agent → File attached to that agent

• Information for the entire company → Workspace Context

If multiple people are migrating, don't have everyone maintain their own version of company facts. Choose someone to manage the Workspace Context.

Skills can be shared across the workspace or kept private. Make a deliberate choice for each skill.

Tip: Add a date and an owner to every document.

Step 8: Verify with real work

Successfully transferring files and instructions doesn't automatically mean the new agent will work like your previous LLM.

Choose five real tasks from the past few months for which you still have good results from your previous LLM. Run those same tasks with the awork Agent and compare the results.

Pay particular attention to workflows that depend on:

• multiple skills or reference files

• important decision paths

• external tools or data sources

• fixed output format requirements

If the results are worse, you have a concrete indication of what information is still missing. Add that specific information rather than transferring more files or materials just in case.

Then expect to make small adjustments over the next two to four weeks. Keep a list of the changes you have to make repeatedly, and regularly turn recurring rules into consistent instructions in an agent's System Prompt or the Workspace Context.

Alternative: Use the full export from ChatGPT or Claude

Export from ChatGPT or Claude

1. In Claude, go to Settings → Privacy → Export data, or in ChatGPT, go to Settings → Data Controls → Export Data.
   Both providers will email you a download link. (Keep in mind that the link expires.)

2. If you're on a Team or Enterprise plan, contact your administrator, as individual members usually can't start an export.

3. Unzip the download. You'll find one or more large JSON files containing all your conversations. They won't be sorted. Some exports include a manifest file that contains only download links. Use the actual conversation data.

4. Find the conversations you need and process them individually using the handover template from Step 4. Don't simply summarize the entire export.

5. Keep the full export as an archive outside your systems and set a deletion date.

Important notes

So make sure that:

• The Workspace Context remains a shared source

• Personal preferences stay separate

• Skills are deliberately shared or kept private

• Documents have an owner and a date

• You check the new awork Agent's results using real tasks

• You regularly turn recurring corrections into consistent rules

Handling company data

An export may contain customer names, contract details, internal decisions, and personal data about colleagues and third parties.

Transferring this information to another system is therefore a data transfer. If you're using a company account, this isn't a decision individual employees should make on their own.

FAQ

Can I import my ChatGPT or Claude conversations directly into awork?

No. You can't import conversations directly. Transferring raw chat data wouldn't be useful anyway, since it may also contain discarded, outdated, or ambiguous content.

Can I export only specific conversations?

No. Providers generally export your entire account history as a bundle of JSON files. You can't select individual conversations during export.

Where should company knowledge go?

Company-wide facts belong in the Workspace Context. Content that only one agent needs should be added as files to that agent.

How long does the migration take?

According to the article, the quick option takes about 30 to 60 minutes and delivers roughly 80% of the result. After that, you should typically expect another two to four weeks of minor adjustments.

What should I do with the full export?

Use it only if you need information from old conversations that isn't available anywhere else. Process the conversations you need selectively, then keep the raw export as an archive outside your systems.