One AI data scientist. Every capability.
Models, voices, LLMs, dashboards and Descartes. Plans include a metered usage balance.
- $5 one-time welcome balance, valid for 10 days; subscribe when spent or expired
- Try every capability: models, dashboards, voice, Descartes
- Every experiment runs isolated, on its own compute
- GPU training for voices and LLMs, paid from the balance
- 2 runs at once
- Community support
- $40 included usage balance / month, separate from the $10 subscription fee
- Pay for what you use: light work costs less
- Isolated GPU compute for voices, speech recognition & LLM fine-tuning
- 3 runs at once, each isolated
- Descartes company brain, up to 5 brains
- Deploy models as APIs and build apps on them
- Sell your models on the Marketplace
- Scheduled retraining
- Snowflake, BigQuery, Redshift, S3
- $150 included usage balance / month
- Everything in Pro
- 5 runs at once
- Pick the AI model for every task
- Desktop app: macOS, Windows, Linux
- Higher API rate limits
- Model monitoring
- Premium support
- Everything in Pro+, $500 shared usage / month
- 8 runs at once across the team
- Shared workspaces & datasets
- Member roles & permissions
- Centralized billing & seats
- Usage analytics & audit
- Everything in Team, usage and parallel runs sized to your organization
- Private cloud, VPC or on-premises
- White label: your brand on your domain
- Seat licences by company email domain
- Call intelligence pushed to your CRM
- Custom connectors to your systems
- Dedicated support & SLA
- Annual invoicing & procurement support
Everything OctOpus does, metered by the work.
Churn, risk, demand and prices, trained and tested on data held back.
Built from your data in seconds, in your house or journal style, with Present.
Clean, reshape and query files of any size, in plain words.
Ask for data: OctOpus finds it, loads it and keeps its lineage.
Your documents, procedures and goals behind every answer. Up to 5 brains, on or off per project.
Train a voice from your recordings and speak any text with it.
OctOpus ASR transcribes calls; each call is coded, measured and exported.
Adapt open language models on your own examples.
PDFs, Word files and images read, summarised, their tables turned into data.
Serve models with scoped keys; the CLI, SDK and Claude Code plugin.
Verify a proof with the OctOpus instance that issued it. Inspect an example.
Databases, warehouses and cloud storage, read safely with masking.
Pay for what you use, at published rates.
Subscription prices and usage balances are separate. For example, Pro costs $10 per month and includes $40 of OctOpus usage at the published rates:
$4 per million input tokens · $18 per million output tokens processed on your behalf.
Operations have published estimates; measured usage and the applicable task limit determine the final charge.
The usage balance pays for OctOpus work and is not a cash payout or a provider account balance.
Reading your existing work is free. Analysis counts vary with the task and compute required.
The full machine-readable rate card is public at /api/pricing/rates.
Procurement, security and private deployment.
For organizations rolling OctOpus out across a division or a regulated environment. We support real review cycles: data handling, subprocessors, private deployment, invoicing. Not a generic PDF.