> For the complete documentation index, see [llms.txt](https://docs.neohive.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.neohive.ai/migration.md).

# Migrating away from Markdown

Import your CLAUDE.md, AGENTS.md, and rules files into NeoHive: searchable, shared, and always current.

If you've been giving your agent context through markdown files (`CLAUDE.md`, `AGENTS.md`, `.cursor/rules`, and the like), you can import all of it into NeoHive. Once it's in, that knowledge becomes searchable by meaning, shared across your team, and it stays current alongside everything else your agent learns.

## Why migrate

Static context files work, but they hit limits as they grow:

* **They're always loaded in full.** The bigger they get, the more of your agent's context window they eat, relevant or not.
* **They're not searchable by meaning.** Your agent reads them top to bottom; it can't pull just the one convention that matters right now.
* **They don't update themselves.** When practices change, someone has to remember to edit the file.

NeoHive stores the same knowledge but retrieves it selectively: only what's relevant to the task in front of you.

## How to migrate

Run the migration skill in your agent:

```
/neohive:migrate-memory
```

1. **It scans your local context files**: `CLAUDE.md`, `AGENTS.md`, `.cursor/rules`, `.codex/rules`, and anything in `.claude/rules`.
2. **It classifies each entry** into a memory type (`directive`, `convention`, `decision`, `insight`, and so on), so NeoHive can weigh and surface it appropriately.
3. **It stores them in your project's knowledge hive**, where every agent on the project can retrieve them.

Your original files are never modified or deleted. Keep them as a reference, or remove them once you're happy with the migration.

### Before and after

A line in your `CLAUDE.md`:

> Always use idempotency keys on payment POST requests.

becomes a stored `directive` in NeoHive, one that surfaces automatically the next time anyone touches the payments code, without loading the whole file.

## After migrating

Nothing changes about how you work. Your agent pulls the migrated knowledge the same way it pulls anything else, when it's relevant.

If something migrated wrong or goes stale, just tell your agent:

* That convention about semicolons is out of date. We switched to no-semicolons last month.

It'll update or replace it in NeoHive.

{% hint style="info" %}
Migrating markdown *notes* (an Obsidian vault, design docs, runbooks) works a little differently: those go into a **document hive** rather than through the migration skill. Upload them through the dashboard (a dedicated guide is on the way).
{% endhint %}

## What about new team members?

When someone new joins and connects their agent to the same project, they inherit everything that's been stored, including the knowledge you migrated. No per-person `CLAUDE.md` setup required.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.neohive.ai/migration.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.