Why is the RAG tier free?

Because creating an organisation is how anyone reaches the product at all, and a card wall there would stop evaluation before it starts. A new organisation's first workspace is always RAG, and RAG costs us vector storage rather than model calls. The GRaaS tier runs a language-model call per shard per document, which is why that one needs a payment method.

What exactly is a workspace?

The tenancy boundary. A workspace scopes all of your data and all of your API keys — documents, entities, the graph, the ontology, everything. An organisation can hold many workspaces, and the GRaaS entitlement is granted at the organisation level, so one payment method unlocks the graph tier for every workspace you own.

What happens to my existing documents when I upgrade to GRaaS?

They stay indexed and searchable, but they are not retroactively extracted into the graph. Upgrading changes what happens to documents ingested from that point on. If you want your existing corpus in the graph, re-POST it — ingest is an idempotent upsert on the document's external ID, so re-sending is safe and replaces rather than duplicates.

Can I downgrade from GRaaS back to RAG?

No. The upgrade is one-way per workspace. If you want to compare the two, create a second workspace on the RAG tier and ingest the same content — the search history in the console can also run any past query against RAG-only retrieval and show you both results side by side.

How are the quotas counted?

Ingests per day and searches per minute are counted per workspace against the rate plan your organisation is on. Document limits are counted as documents currently stored, not documents ever sent. Deleting a document frees its slot.

Are you SOC 2 certified?

Not yet, and we would rather say so than imply otherwise. Our security page describes exactly what is in place today — workspace isolation, scoped keys with rotation, single sign-on through WorkOS, session revocation — and what is still on the roadmap.

Will you sign a data processing agreement?

Yes. Raise it during the demo conversation and we will get it in front of you before you send any production content.

Do I need to bring my own model?

No, and you currently cannot. Extraction, embedding and answer synthesis run on models we select and pay for, which is part of what the GRaaS tier covers. If a specific model or region is a requirement for you, that is an Enterprise conversation.

Pricing

Free until it's worth paying for.

Ten thousand documents, real answers with citations, no card and no call. You move up when your content spans enough places that connecting it starts to matter.

Book a demo Start free

no card for the free tier
graph entitlement is org-wide
upgrade is one-way

Free
RAG tier
$0
For finding out whether this works on your content. Real answers with citations over ten thousand documents, no card, no call. A new organisation always starts here.
- 10,000 documents
- 1,000 ingests per day
- 30 searches per minute
- Cited answers with enforced grounding
- Search Playground and search history
- MCP server
Start free
Starter
GRaaS tier
Talk to us
For a first production use case. Everything in Free, plus the connections between your documents and a vocabulary your team controls.
- 5,000 documents
- 2 concurrent extractions
- Knowledge graph and three-lane retrieval
- Versioned ontology, impact preview, drift
- Graph visualizer and entity quality
- Corpus health
Book a demo
Growth
most chosen
GRaaS tier
Talk to us
For content that has outgrown one system and a team that actively curates it. The point at which connecting your sources starts paying for itself.
- 50,000 documents
- 10,000 ingests per day
- 120 searches per minute
- 8 concurrent extractions
- Everything in Starter
- Batches of 100 documents
Book a demo
Enterprise
GRaaS tier
Custom
For putting your own customers behind it. Negotiated capacity, a dedicated plan, and our engineers in your integration.
- 500,000 documents and up
- Unlimited ingest and search rates
- Unlimited concurrent extractions
- Per-organisation capacity ceilings
- Hands-on integration support
- Everything in Growth
Talk to sales

Graph-tier prices are not published yet, and nothing on this site takes a payment today. Tell us what you are building and we will scope it with you — the quota numbers below are real and already enforced by the plan you would be on.

Compare

Every plan, every limit.

These are the quotas the platform actually enforces, not marketing rounding. A test in our own build fails if this table and the running system disagree.

Feature and quota comparison across the four plans
Feature	FreeRAG tier	StarterGRaaS tier	GrowthGRaaS tier	EnterpriseGRaaS tier
Capacity
Documents	10,000	5,000	50,000	500,000+
Ingests per day	1,000	1,000	10,000	Unlimited
Searches per minute	30	30	120	Unlimited
Documents per batch	25	25	100	Unlimited
Concurrent extractions	—	2	8	Unlimited
Retrieval
Hybrid dense + keyword search	Included	Included	Included	Included
Cited answers with enforced grounding	Included	Included	Included	Included
Graph expansion lane	Not included	Included	Included	Included
Evidence lane, ranked by corroboration	Not included	Included	Included	Included
Per-step search diagnostics	Included	Included	Included	Included
Graph and governance
Knowledge graph	Not included	Included	Included	Included
Multi-source provenance	Not included	Included	Included	Included
Versioned ontology with impact preview	Not included	Included	Included	Included
Drift detection	Not included	Included	Included	Included
Graph visualizer	Not included	Included	Included	Included
Entity quality and merge queue	Not included	Included	Included	Included
Corpus health	Not included	Included	Included	Included
Platform
MCP server, 12 tools	Included	Included	Included	Included
Scoped API keys with rotation	Included	Included	Included	Included
Search history and retrace	Included	Included	Included	Included
Multiple workspaces	Included	Included	Included	Included
Single sign-on	Included	Included	Included	Included

Free

RAG tier

Documents: 10,000
Ingests per day: 1,000
Searches per minute: 30
Documents per batch: 25
Concurrent extractions: —
Hybrid dense + keyword search: Included
Cited answers with enforced grounding: Included
Graph expansion lane: Not included
Evidence lane, ranked by corroboration: Not included
Per-step search diagnostics: Included
Knowledge graph: Not included
Multi-source provenance: Not included
Versioned ontology with impact preview: Not included
Drift detection: Not included
Graph visualizer: Not included
Entity quality and merge queue: Not included
Corpus health: Not included
MCP server, 12 tools: Included
Scoped API keys with rotation: Included
Search history and retrace: Included
Multiple workspaces: Included
Single sign-on: Included

Starter

GRaaS tier

Documents: 5,000
Ingests per day: 1,000
Searches per minute: 30
Documents per batch: 25
Concurrent extractions: 2
Hybrid dense + keyword search: Included
Cited answers with enforced grounding: Included
Graph expansion lane: Included
Evidence lane, ranked by corroboration: Included
Per-step search diagnostics: Included
Knowledge graph: Included
Multi-source provenance: Included
Versioned ontology with impact preview: Included
Drift detection: Included
Graph visualizer: Included
Entity quality and merge queue: Included
Corpus health: Included
MCP server, 12 tools: Included
Scoped API keys with rotation: Included
Search history and retrace: Included
Multiple workspaces: Included
Single sign-on: Included

Growth

GRaaS tier

Documents: 50,000
Ingests per day: 10,000
Searches per minute: 120
Documents per batch: 100
Concurrent extractions: 8
Hybrid dense + keyword search: Included
Cited answers with enforced grounding: Included
Graph expansion lane: Included
Evidence lane, ranked by corroboration: Included
Per-step search diagnostics: Included
Knowledge graph: Included
Multi-source provenance: Included
Versioned ontology with impact preview: Included
Drift detection: Included
Graph visualizer: Included
Entity quality and merge queue: Included
Corpus health: Included
MCP server, 12 tools: Included
Scoped API keys with rotation: Included
Search history and retrace: Included
Multiple workspaces: Included
Single sign-on: Included

Enterprise

GRaaS tier

Documents: 500,000+
Ingests per day: Unlimited
Searches per minute: Unlimited
Documents per batch: Unlimited
Concurrent extractions: Unlimited
Hybrid dense + keyword search: Included
Cited answers with enforced grounding: Included
Graph expansion lane: Included
Evidence lane, ranked by corroboration: Included
Per-step search diagnostics: Included
Knowledge graph: Included
Multi-source provenance: Included
Versioned ontology with impact preview: Included
Drift detection: Included
Graph visualizer: Included
Entity quality and merge queue: Included
Corpus health: Included
MCP server, 12 tools: Included
Scoped API keys with rotation: Included
Search history and retrace: Included
Multiple workspaces: Included
Single sign-on: Included

The real difference

Why the graph tier costs anything at all.

Worth understanding, because it explains the shape of the pricing rather than just the numbers.

RAG costs us storage

Chunking and embedding are cheap and happen once. Serving a vector search is cheap. That is why the free tier can be genuinely free rather than a two-week trial.

The graph costs us model calls

Extraction runs a language-model call per shard per document. It is not expensive per document — roughly a fifth of a cent for ten thousand — but it is unbounded if anyone can point a firehose at it, which is why the graph tier needs a payment method on file.

Entitlement is per organisation

One card, or one arrangement with us, unlocks the graph tier for every workspace your organisation owns. You do not pay per workspace to evaluate it.

FAQ

Frequently asked

Billing, quotas, and the two things people are most surprised by.

Start on the free tier.

Create an organisation, get a workspace, issue a key, POST a document. No card at any point.

no card required
first workspace is free
one POST to ingest

Talk to us Start free