Pricing
Self-host free or let us manage it
Give your agents memory that learns. Self-host free with a single Docker command, use Hindsight Cloud — pay-as-you-go on token usage, no fixed monthly fee — or run Hindsight Enterprise in your cloud or ours. Hindsight Models are priced per million tokens.
Self-hosted
Run Hindsight 1.0 on your own infrastructure with a single Docker command. MIT licensed, no restrictions.
- MIT open source license
- Single Docker command to deploy
- All four memory networks
- Retain, Recall, and Reflect APIs
- MCP server built in
- Embedded PostgreSQL
- Community support via GitHub
Hindsight Cloud
Managed Hindsight with zero infrastructure. Billed on token usage — no fixed monthly fee, no seat pricing. Start free, scale with your agents.
- Everything in Self-hosted
- Usage-based billing on tokens
- Start for Free
- Managed infrastructure
- Automatic scaling
- Daily backups
- Dashboard and usage analytics
- Team collaboration
- Support SLA available
- 99.9% uptime guarantee
Enterprise
Hindsight in a private cloud, your own AWS, Google Cloud or Azure account, or on premises, with enterprise security and contracted support.
- Everything in Cloud
- Private cloud, BYOC or on-premises
- Single-tenant deployment
- Private networking
- Hindsight Memlake for large memory volumes
- SSO and enforced MFA
- Memory Defense
- Audit logs, webhooks and SIEM
- Custom SLA (up to 99.95%)
- Shared Slack channel with our team
- Up to 24×7 support, 30-minute response SLA
- Invoiced billing
Hindsight Models
Decision, rerank and embedding models trained for Hindsight. Point your self-hosted Hindsight at Hindsight Inference and pay per million input tokens with output free, or run the open weights yourself at no cost.
Decision model
hs-decide-1-flash
- Decisions API (TypeSafe System One-compatible)
- Output tokens free
- 0.8B parameters, open weights (Apache 2.0)
Reranker
hs-rerank-1
- Reranker API (Cohere Rerank-compatible)
- Output tokens free
- 114M parameters, open weights (MIT)
Embeddings
hs-embed-1
- Embeddings API (OpenAI Embeddings-compatible)
- Output tokens free
- 112M parameters, open weights (MIT)
Billed from the same credit balance as Hindsight Cloud, with the same login and organization. New accounts start with $5 of free credit. Inference docs
Compare plans
| Feature | Self-hosted | Cloud | Enterprise |
|---|---|---|---|
| Full learning system (4 memory networks) | |||
| Store, retrieve, and learn from experience | |||
| MCP Server | |||
| Python SDK | |||
| REST API | |||
| Managed Infrastructure | — | ||
| Automatic Scaling | — | ||
| Daily Backups | — | ||
| Dashboard & Analytics | — | ||
| Team Collaboration | — | ||
| Private cloud, BYOC or on-premises | — | — | |
| Single-tenant deployment | — | — | |
| Private networking | — | — | |
| SSO and enforced MFA | — | — | |
| Memory Defense | — | — | |
| Audit logs, webhooks and SIEM | — | — | |
| Data privacy (GDPR) controls | — | — | |
| Invoiced billing | — | — | |
| Shared Slack channel | — | — | |
| Uptime SLA | N/A | 99.9% | Up to 99.95% |
| Support | Community | 12×5 email | Up to 24×7 |
| Response SLA | N/A | 12 hours | 30 minutes |
Frequently asked questions
- Is Hindsight really free to self-host? +
- Yes. Hindsight is MIT licensed with no usage limits, no telemetry, and no restrictions. Run it in production on your own infrastructure at no cost.
- What is the difference between self-hosted and Hindsight Cloud? +
- Self-hosted means you run Hindsight on your own servers using Docker. Hindsight Cloud is our fully managed service — we handle infrastructure, scaling, backups, and updates so you can focus on building.
- Can I migrate from self-hosted to Cloud later? +
- Yes. Contact us for assistance migrating your memory data from a self-hosted instance to Hindsight Cloud.
- How does Hindsight Cloud pricing work? +
- Hindsight Cloud is pay-as-you-go. There is no fixed monthly fee and no per-seat pricing. Each operation is billed in the unit that matches how it works — per million tokens for token-driven operations, per call for operations that run a fixed workload:
- Retain — storing memories
- Recall — retrieving memories
- Reflect — synthesizing patterns from experience
- Iris Extract — structured extraction
- Mental Model — retrieve and refresh operations
- Storage — long-term memory held past its first 30 days
- How much do operations cost? +
- Retain (store memories) — $10.00 per million tokens
- Recall (retrieve memories) — $0.75 per million tokens
- Reflect (synthesize patterns) — $0.05 per call
- Iris Extract (structured extraction) — $7.50 per million tokens
- Mental Model Retrieve — $0.25 per million tokens
- Mental Model Refresh — $0.05 per call
- Storage — $0.25 per million tokens stored per month
- How much do Hindsight Models cost? +
- On Hindsight Inference, the Hindsight Models are billed per million input tokens from the same credit balance as Hindsight Cloud. Output is free, and the open weights are free to run yourself.
- hs-decide-1-flash (Decisions API) — $0.02 per million tokens
- hs-rerank-1 (Reranker API) — $0.015 per million tokens, or $0.02 with hs-decide-1-flash to also drop irrelevant documents
- hs-embed-1 (Embeddings API) — $0.005 per million tokens
- What is the difference between Hindsight Cloud and Enterprise? +
- Hindsight Cloud is our shared, self-serve service. Hindsight Enterprise adds enterprise security and contracted support, and can run as a dedicated single-tenant deployment in our cloud, in your own AWS, Google Cloud or Azure account, or on premises. Learn about Enterprise.
- Do you offer volume discounts? +
- Yes. If you expect significant token volume, contact us to discuss custom pricing.
- Is there a free tier? +
- You get free credits to get started, so you can build and test a real agent before any charges kick in.
Need help choosing?
Our team is here to help you find the right plan.
Docs: Open source · Enterprise
Looking for the Vectorize RAG platform? platform.vectorize.io →