Vellum
A persistent personal AI assistant with memory, connected tools, hosted plans, and a self-hosting option.
About Vellum
Vellum is a personal AI assistant that runs in the background, remembers preferences, and connects to services including Gmail, Slack, and Telegram. The assistant has its own identity, accounts, memory, and tools, so it can handle work outside a browser chat. The hosted Base plan has no monthly plan fee. It includes a fixed Small machine with 1 vCPU, 2 GiB RAM, and 4 GiB storage, with usage credits billed as they are consumed. Pro starts at $50 per month for the documented configuration and adds configurable compute, more storage, an assistant email address, and a subdomain. The open-source assistant can also run on your own hardware, a VPS, or a private cloud. Self-hosting avoids Vellum platform and machine-tier fees, but the operator becomes responsible for data, keys, infrastructure, and uptime. Vellum suits recurring connected workflows where a persistent assistant is useful. Buyers still need to account for usage credits and any extra compute or storage.
Best for
Best use cases
Run inbox triage and messaging tasks through connected services
Keep calendar, meeting preparation, and delegated follow-up moving in the background
Build and persist a custom TypeScript skill for an uncovered workflow
Compare before choosing
Base removes the monthly plan fee but fixes compute and storage at Small and 4 GiB
Pro adds configurable compute, storage, email, and a subdomain, while credits remain separate
Self-hosting removes Vellum infrastructure fees but transfers operations and uptime responsibility
Features
Email and messaging operations
Connect Gmail, Slack, or Telegram so the assistant can read, search, send, archive, and react through provider-specific tools.
Background personal workflows
Keep recurring work moving while away, such as inbox triage, calendar coordination, meeting preparation, and delegated follow-up.
Custom skill automation
Create, test, save, load, and remove TypeScript-based skills at runtime when the existing tool set does not cover a workflow.
Pros
- Base has no monthly plan fee, so hosted adoption can start with usage-based credit costs
- Gmail, Slack, and Telegram integrations support practical messaging workflows instead of chat alone
- The open-source self-hosting option gives operators control over data, keys, infrastructure, and uptime
Cons
- Usage credits are billed separately and can rise with LLM inference, search, images, and paid APIs
- Base is fixed at 1 vCPU, 2 GiB RAM, and 4 GiB storage, limiting heavier hosted workloads
- Self-hosting removes platform fees but requires the user to operate infrastructure and manage uptime
Pricing Plans
Base
- Pay-as-you-go usage credits
- Small machine: 1 vCPU and 2 GiB RAM
- 4 GiB storage
Pro
- $10/month platform fee
- Medium processing power: +$35/month
- 10 GiB storage: +$5/month
- Usage credits billed separately
- Email and subdomain included
Self-hosted
- Deploy on your own hardware, VPS, or private cloud
- No Vellum platform fee or machine-tier charge
- You manage data, keys, infrastructure, and uptime
* Pricing information is for reference only. Please refer to the latest announcement on Vellum's official website.