Predictable pricing

Generous free tier. One cheap entry plan. A team plan that doesn't double your bill. Usage-based overages on Pro and Team. Cancel anytime.

No surprise overages — usage alerts before you hit a limit. Cancel anytime.

free

$0

Build & evaluate

  • Stored vectors50K
  • Queries / mo5K
  • Answers / mo200
  • …on our model key200
  • Reranks on our key / mo—
  • Embed tokens / mo5M
  • Storage1 GB
  • Projects2
  • Indexes3
  • Members1
  • Rate limit / min60
  • Hybrid search
  • Reranking
  • BYO model keys
  • RBAC roles
  • Overage on caps
pro

$25/mo

Ship to production

  • Stored vectors500K
  • Queries / mo500K
  • Answers / mo50K
  • …on our model key10K
  • Reranks on our key / mo2K
  • Embed tokens / mo50M
  • Storage50 GB
  • Projects3
  • Indexesunlimited
  • Members3
  • Rate limit / min600
  • Hybrid search
  • Reranking
  • BYO model keys
  • RBAC roles
  • Overage on caps
team

$99/mo

Scale with a team

  • Stored vectors2.5M
  • Queries / mo2M
  • Answers / mo250K
  • …on our model key50K
  • Reranks on our key / mo10K
  • Embed tokens / mo250M
  • Storage250 GB
  • Projectsunlimited
  • Indexesunlimited
  • Membersunlimited
  • Rate limit / min3K
  • Hybrid search
  • Reranking
  • BYO model keys
  • RBAC roles
  • Overage on caps
enterprise

Custom

Compliance & scale

  • Stored vectors∞
  • Queries / mo∞
  • Answers / mo∞
  • …on our model key∞
  • Reranks on our key / mo∞
  • Embed tokens / mo∞
  • Storage∞
  • Projectsunlimited
  • Indexesunlimited
  • Membersunlimited
  • Rate limit / min10K
  • Hybrid search
  • Reranking
  • BYO model keys
  • RBAC roles
  • Overage on caps

Overage rates (Pro & Team)

Billed monthly via Stripe metered usage. Free hard-caps instead of charging.

Stored vectors$1.00 / 100K chunks per month
Queries$0.50 per 1,000 queries
Answers$1.00 per 1,000 answers
Answers on our model key+$2.00 per 1,000 answers
Reranks on our key$5.00 per 1,000 reranks
Embed tokens (platform key)$0.04 per 1M tokens
Doc storage$0.10 per GB per month

With your own model keys, answers and reranks skip the “on our key” meters — you pay your provider directly.

Pricing FAQ

What counts as a "stored vector"?

Every chunk you ingest produces one vector. We meter the live count across all indexes in your org.

What is the difference between a query and an answer?

A query returns ranked chunks. An answer runs the query then asks an LLM to write a cited response with the retrieved chunks as grounding.

What happens when I hit a cap?

Free hard-blocks at 100% with a friendly upgrade prompt. Pro and Team allow overage, billed via Stripe metered usage.

Which models can I use?

Answers use Claude Haiku 5.5 by default on our model key, within your plan's allowance. To use Claude Sonnet or Opus, or GPT-4o, add your own Anthropic or OpenAI key to the project.

Do BYO model keys reduce cost?

Yes. Answers and reranks that run on our keys have their own monthly allowance. With your own Anthropic, OpenAI or Cohere key they only count toward your plan's answer and query allowances, and you pay your provider directly for the tokens.

Can I self-host?

Self-hosting is available on Enterprise plans. Contact us to discuss a BYOC or on-premise deployment.

No credit card to start

Start free. Upgrade when you ship.