assistant-migration

Migrate memory, identity, skills, and integrations from other AI assistants into Vellum.

1.2k|166|Updated Feb 7, 2026
One-click install
npx skills add https://github.com/vellum-ai/vellum-assistant --skill assistant-migration
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: assistant-migration
Source: https://github.com/vellum-ai/vellum-assistant/tree/main/skills/assistant-migration
Command: npx skills add https://github.com/vellum-ai/vellum-assistant --skill assistant-migration

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Switching from ChatGPT, Claude, OpenClaw, Hermes, or Manus to a new assistant usually means losing years of memory, custom instructions, skills, and integrations. This Skill inspects source assistant exports and workspaces, builds a reviewed inventory, and maps everything safely portable into Vellum primitives.

Core Features & Use Cases

  • Unified multi-source inventory: Combines exports from several assistants into one deduplicated checklist with source attribution, confidence levels, and per-item port/review/re-setup/disregard actions.
  • Deterministic memory extraction: Bundled parsers introspect ChatGPT export ZIPs and Hermes/OpenClaw SQLite memory snapshots without assuming schemas, redact credential-shaped values, and emit review candidates for staged v3 memory ingestion.
  • Per-assistant reconnaissance references: ChatGPT, Claude, Hermes, and OpenClaw references provide data-directory locations, tar bundling recipes with secret exclusions, and post-import credential rebind checklists.
  • Use Case: A user moving from ChatGPT plus a self-hosted Hermes agent attaches both archives, reviews a single combined inventory, imports approved memories via assistant memory ingest, recreates skills, and rebinds OAuth credentials through the vault.

Quick Start

Ask the assistant to migrate your data from ChatGPT, Claude, OpenClaw, or Hermes into Vellum and attach your export archive to the conversation.

Frequently Asked Questions about assistant-migration

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I migrate from ChatGPT to Vellum?

Request the official ChatGPT data export from Settings, then attach the ZIP to the conversation. Conversation history is handled by the chatgpt-import skill, while custom instructions and saved memories are extracted by the bundled parser and reviewed before import.

How do I import Hermes or OpenClaw memory into Vellum?

Create a consistent snapshot of memory.db using sqlite3 .backup, then run the bundled parse-agent-memory-db.ts script with the appropriate --source flag. It introspects the schema, redacts credential-shaped values, and emits review candidates for staged ingestion.

Can I migrate from multiple assistants at once?

Yes, the skill builds one unified inventory across all sources with a source attribution column. Duplicate facts are collapsed into single Vellum items with provenance notes, and conflicts are surfaced for your decision.

Does the migration import API keys and OAuth tokens?

No, secrets are never imported from exports, logs, or config dumps. All credentials are reconnected through Vellum's credential vault, OAuth connect flows, or secure prompts after the migration completes.

What happens to data that cannot be safely ported?

Opaque items like vector stores, caches, binary databases, and foreign runtime state are marked for review or rebuild rather than copied. The final migration report lists what was ported, needs review, needs re-setup, or was disregarded.