Sistava vs Vellum
AI Personal Assistants
Sistava vs Vellum
Vellum started in 2023 as a Y Combinator-backed platform for building and evaluating enterprise LLM applications, then relaunched in 2026 as a personal AI assistant under the tagline "Your Personal Intelligence." It runs across web, macOS, iOS, CLI, email, Telegram, and Slack, holds persistent memory, and gets its own identity so it can act on your behalf. The open-source MIT runtime can deploy to Vellum Cloud or your own infrastructure. It is currently in early access via a waitlist, with no paid pricing announced yet.
Profile
- Pricing
- Early access via waitlist. Free one-click install planned for macOS. No paid tiers announced yet.
- Founded
- 2023, New York, as a Y Combinator-backed enterprise AI dev platform; relaunched in 2026 as a personal AI assistant under "Vellum: Your Personal Intelligence"
- Funding
- $25.5M+ raised across seed and Series A ($20M Series A led by Leaders Fund, July 2025)
- Focus
- Personal AI assistant with persistent memory that browses the web, manages files and calendars, and acts under its own identity (a dedicated email, GitHub account, and Slack handle)
Why Sistava
- Vellum is a single personal assistant for one person. Sistava is built to scale from one AI employee to a full team of specialists (sales, marketing, support, ops) under a leader who delegates.
- Vellum is in early access with a waitlist and no announced pricing. Sistava is live today with published plans (Free + $49/$99/$199/$399 per month) you can sign up for right now.
- Honest concession: Vellum's identity model, giving the assistant its own email, GitHub account, and Slack handle so it acts as a real participant rather than a tool bolted onto your accounts, is a genuinely interesting design choice we have not shipped.
Feature Comparison
| Feature | Vellum | Sistava |
|---|---|---|
| Early access waitlist | Live, self-serve signup | |
| Single personal assistant | Unlimited AI employees, team hierarchy | |
| ✓ | ✗ | |
| Persistent memory (single-assistant level) | 7-layer memory system | |
| Yes, MIT runtime | ✗ | |
| Not announced (early access) | Free + $49/mo |