OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Memory & RAG

Vellum Assistant

Vellum Assistant is an MIT-licensed TypeScript personal AI agent that stores multiple forms of user memory, learns working preferences and performs approved actions across applications and communication channels.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Vellum Assistant — OpenViST project profile in Agent Memory & RAG.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when continuity, preference-aware assistance or approved actions across several apps matter more than a single isolated exchange.

Who it is for

  • People seeking a persistent AI assistant that retains preferences and work context.
  • Developers and advanced users interested in a TypeScript, open-source assistant with CLI and self-hosting options.
  • Users whose workflows span supported applications, messaging channels or computer-use tasks.

Documented uses

  • Retain and retrieve personal context across conversations.
  • Learn how a user communicates and works over time.
  • Identify unfinished or upcoming work and send proactive messages.
  • Perform approved file, command, browser and application actions.
  • Extend the assistant with runtime skills and tools.
Before you choose

Requirements & limits

The repository is public and specifies an MIT license. It documents managed and local modes, desktop installation, and CLI installation from Bun or source. The desktop app is identified as the primary focus.

  • Its eight separate memory types provide a more structured memory model than a simple notes-based assistant.
  • Hybrid dense and sparse retrieval, structured extraction, source attribution and deduplication shape how remembered information is organized.
  • Managed and self-hosted operation support different deployment preferences.
  • Computer-use actions are sandboxed and permission-gated, with approval durations described as once, ten minutes or always.
  • The project references Anthropic and OpenAI ecosystems, while its implementation language is TypeScript.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up Vellum Assistant through the documented CLI path

Install the CLI or repository version and create an assistant with the documented hatch command.

Before you start

  • A terminal-based workflow.
  • Bun for the global installation route, or a cloned copy of the public repository for installation from source.

Steps

  1. Install the CLI globally with the documented Bun command, or clone the repository and run its setup script.

  2. Reload the shell configuration after the source setup path as shown in the repository instructions.

  3. Run vellum hatch to create the assistant.

  4. Use vellum wake to start services when needed.

Expected outcome

An assistant is created and its services can be started through the CLI.

Guide limitations

  • The repository describes the CLI as working but says the desktop app is the primary focus.
  • The guide documents setup and service commands, not the result of a particular assistant task.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
4 October 2026
Original website title
vellum-ai/vellum-assistant · Official homepage title · retained in its source language
How we review sourcesSuggest a correction