Run multimodal AI in production
Unified Storage, Orchestration and Retrieval.
Community
For trying the cloud, free
- 1 free organization
- 1 cloud database
- Unlimited tables
- $30/mo compute credit
- 1 TB Media Store (hard limit)
- 10 GB database storage (hard limit)
- Community support
Pro
For production apps and teams
- Everything in Community, plus:
- 3 cloud databases
- $59/mo compute credit
- $0.40 / GB-mo DB storage above 10 GB
- Bring your own bucket for media storage
- Custom domains for Deploy endpoints
- Email support
Enterprise
Custom deployment at scale
- Everything in Pro, plus:
- Single-tenant VPC & BYOC
- Dedicated database and runtime pools
- Bring-your-own GPU (in development)
- Dedicated support & SLAs
- Volume discounts
Usage-based CPU, GPU, and storage after monthly credits and included limits. See rates below.
Self-Hosted
Open SourceThe full engine — unified multimodal storage, incremental pipelines, embeddings, and semantic search. Run it anywhere under Apache 2.0.
How billing works
One platform, one usage wallet
Queries, vector search, endpoints, workflows, table compute, and GPU all roll into compute-hours.
Database
Cloud-hosted multimodal tables with separate meters for Media Store (home-bucket media) and database disk (rows, embeddings, metadata). One database on Community, unlimited on Pro.
Compute-hours
One wallet for table execution, queries, vector search, workflows, endpoints, CPU, and GPU. Database reads and API responses are covered by compute-hours, not a separate egress bill.
Storage
Media Store: 1 TB hard limit on every plan. Database storage: 10 GB hard limit on Community; 10 GB included on Pro with usage-based overage above. BYOB media stays free.
Compare plans
Databases, endpoints, usage, and support — side by side.
| Feature | Community | Pro | Enterprise |
|---|---|---|---|
| Cloud Databases | |||
| Organizations per account | 1 free | Unlimited (paid) | Unlimited |
| Cloud databases per org | 1 (Preview) | 3 | Unlimited |
| Tables per database | Unlimited | Unlimited | Unlimited |
| Media Store included | 1 TB (hard limit) | 1 TB (hard limit) | Custom |
| Database storage included | 10 GB (hard limit) | 10 GB | Custom |
| Database storage above included | Upgrade to Pro | $0.40 / GB-mo | Custom |
| Database I/O and egress | Included in compute-hours | Included in compute-hours | Custom |
| Bring your own bucket for media storage | ✗ | ✓ | ✓ |
| SAML SSO | ✗ | Contact us | ✓ |
| Single-tenant VPC deployment | ✗ | ✗ | ✓ |
| Runtime & Compute | |||
| HTTP endpoints (pxt deploy) | Preview | ✓ | ✓ |
| Queries and vector search | Compute-hours | Compute-hours | Custom |
| Workflow and table compute | Compute-hours | Compute-hours | Custom |
| Serverless workers | Scale-to-zero | Scale-to-zero + usage-based warm capacity | Dedicated pools |
| GPU compute | NVIDIA T4 | NVIDIA T4 | Custom capacity |
| Versioned rollbacks | ✓ | ✓ | ✓ |
| Custom domains | ✗ | ✓ | ✓ |
| BYOC (bring your own cloud) | ✗ | ✗ | ✓ |
| Usage & overage | |||
| Compute credit included | $30/mo | $50/mo | Custom |
| CPU compute rate | $0.08 / vCPU-hour | $0.08 / vCPU-hour | Custom |
| GPU compute rate | See rates | See rates | Custom |
| Support | |||
| Community (Discord) | ✓ | ✓ | ✓ |
| Priority support | ✗ | ✓ | ✓ |
| Dedicated support channels | ✗ | ✗ | ✓ |
| SLA guarantees | ✗ | ✗ | ✓ |
| Volume discounts | ✗ | ✗ | ✓ |
Usage rates
Compute and storage pricing
Same rates on Community and Pro. Your monthly compute credit applies first, then pay-as-you-go. Endpoint requests, query counts, and workflow runs are not billed separately — they consume CPU or GPU time.
CPU compute
$0.08 / vCPU-hour
Queries, vector search, Deploy endpoints, workflows, and table compute on CPU workers.
Billed per vCPU-second while workers are active. Scale-to-zero when idle. Database reads and API responses are included — not a separate egress bill.
GPU compute
NVIDIA T4 on managed workers
| GPU class | VRAM | Rate | Typical use |
|---|---|---|---|
| NVIDIA T4 | 16 GB | $0.65 / GPU-hr | Inference, embeddings, small models |
GPU UDFs and model inference draw from the same compute wallet as CPU work. BYO API keys for OpenAI, Anthropic, and other providers are not marked up. Bring-your-own GPU via Runpod, Modal, and other hyperscalers is in development.
Storage
Media Store
Home-bucket media for videos, images, audio, and computed outputs. 1 TB hard limit on every plan. BYOB is free. No egress fees on managed media.
Database storage
Table rows, embeddings, and catalog metadata stored in your cloud database. Community includes 10 GB (hard limit); Pro includes 10 GB, then $0.40 / GB-mo.
Preview pricing subject to change at GA.
Customer voices
Built by teams shipping AI
"Pixeltable's seamless integration into existing vision workflows means our team can rapidly experiment and iterate, delivering high-impact solutions faster than ever."
Adil Mohammad
Founding Engineer, Obvio (fmr. Sr. Deep Learning Engineer, Nvidia)
"Pixeltable has transformed how our ML engineers spend their time... We have a highly multimodal use case, and Pixeltable gives us full visibility into the input data, models, and incremental steps of our system's pipeline."
Denise Kutnick
CEO, Variata (fmr. Director of MLSys , OctoAI)