Start free, on a real engagement
The free plan is not a tour: it runs the whole pipeline, produces the full document library and exports it. Paid plans are about teams, governance and scale rather than unlocking the product.
Enough to run a real engagement through the pipeline and judge the output.
- One organisation with one workspace
- All eight deliverables
- Knowledge uploads and retrieval
- DOCX, PDF and Markdown export
- A monthly spending ceiling that refuses runs rather than billing past it
For a team running client work continuously, with governance a delivery lead can see.
- Multiple workspaces with project isolation
- Role-based access and custom roles
- Usage analytics by provider, model and workspace
- Workspace branding and document format control
- Activity ledger and audit trail
For firms with several delivery teams, their own AI providers and their own controls.
- Multiple organisations and workspaces
- Own provider keys and model selection
- Platform administration: prompts, pricing, document types
- Spend ceilings set per organisation
- Self-hosted deployment
Paid plans are bought in the product, by card, through Stripe or Razorpay — whichever the deployment has switched on. Upgrading, downgrading and renewing are self-serve, and the plan in force, the current period and every payment stay on your organisation’s billing page. Card details are taken by the gateway, never by this application.
Provider usage is billed by whichever AI provider you run on, and metered inside the product against a monthly ceiling you set. Enterprise is quoted per organisation because the shape varies with how many workspaces and providers you need — we would rather discuss it than publish a number that turns out to be wrong for you.
What differs between plans
Everything that produces a document is in every plan. The differences are about teams and control.
| Capability | Free | Pro | Enterprise |
|---|---|---|---|
| Deliverables | |||
| The core pack of eight | Yes | Yes | Yes |
| The full library of 24 types | Yes | Yes | Yes |
| Full-pack generation | Yes | Yes | Yes |
| BA intelligence, quality gate and consistency pass | Yes | Yes | Yes |
| Version history, compare and restore | Yes | Yes | Yes |
| DOCX, PDF and Markdown export | Yes | Yes | Yes |
| Knowledge | |||
| Uploads and retrieval | Yes | Yes | Yes |
| Web research | Yes | Yes | Yes |
| Assistant chat with citations | Yes | Yes | Yes |
| Teams | |||
| Workspaces | One | Several | Unlimited |
| Role-based access | Built-in roles | Built-in and custom | Built-in and custom |
| Invitations | Yes | Yes | Yes |
| Governance | |||
| Spending ceiling | Plan default | Configurable | Per organisation |
| Usage analytics | Basic | Full | Full |
| Activity and audit | Yes | Yes | Yes |
| Platform administration | — | — | Yes |
| Card payment and self-serve plan changes | n/a | Yes | By arrangement |
Before you commit
Anything not answered here, ask us directly — including the awkward ones about data and model providers.
How do we pay?
By card, through Stripe or Razorpay, whichever the deployment has switched on. Upgrading, downgrading and renewing are all done in the product; the plan in force, the current period and every payment are on the organisation's billing page. Card details never touch this application — the gateway takes them on its own page.
What happens when we change plan mid-period?
The new plan takes effect from the moment it is paid for, and the previous subscription is closed out with its own record kept. Nothing is deleted: the history of every plan an organisation has held stays on the billing page.
How is provider usage counted?
Every generation records the provider, the model, the tokens in and out and the cost computed from the rates configured for that model. The organisation's month-to-date total is what the spending ceiling is checked against.
What happens when a spending ceiling is reached?
Generation is refused — for every workspace in the organisation — until the month rolls over or the limit is raised. It is a control, not a warning: nothing keeps billing past it.
Can we use our own AI provider keys?
Yes. Providers are configured per deployment, and the model can be chosen per run. Keys are never returned by any endpoint; only their presence is reported.
Do you train models on our client material?
No. Uploaded documents are extracted, embedded and stored inside the workspace they belong to, and are used to ground that workspace's own generations. Nothing is shared between workspaces.
Can we self-host?
Yes, on the Enterprise plan. The application ships as a container with PostgreSQL and pgvector, and can run entirely inside your own environment.
Evaluate on the free plan first
It runs the full pipeline. If it does not produce documents you would put your name on, no plan above it will either.