LaunchRoundsSubmit Software

Vectorize: Agents that learn from their mistakes

Vectorize is the company behind Hindsight, an open source agent memory layer that stores, recalls and reflects on experience so agents improve over time.

What is Vectorize?

Vectorize builds Hindsight, a memory layer for AI agents. Rather than only storing and retrieving facts, Hindsight turns failed tool calls and user corrections into experiences, and a reflection layer consolidates raw observations into reusable knowledge. Memory is kept per user and persists across sessions, with parallel search returning relevant memories in under 100ms according to the vendor. The layer is model-agnostic, so swapping LLMs does not discard what an agent has learned. Hindsight is MIT licensed and can be self-hosted with a single Docker command, including an embedded PostgreSQL instance and a built-in MCP server. Agents that support MCP get remember, recall and reflect tools, and there are also Python SDK and REST API options. A managed Hindsight Cloud option handles infrastructure, scaling and backups with usage-based billing, and an Enterprise tier adds dedicated deployment, SSO/RBAC and support SLAs. Vectorize also notes a separate RAG platform at platform.vectorize.io.

Focus areas: Agent Memory, Vector Search.

Who is Vectorize for?

  • Engineering teams building LLM agents that need persistent user context
  • Developers using MCP-capable tools such as Claude Code or Cursor
  • Platform teams that want to self-host memory on their own infrastructure

How much does Vectorize cost?

Pricing model: freemium. Always check the vendor’s own pricing page before you buy.

PlanPriceWhat you get
Self-hostedFreeMIT licensed, single Docker command, all four memory networks, Retain/Recall/Reflect APIs, MCP server, embedded PostgreSQL, community support
Hindsight CloudPay as you goUsage-based billing on tokens, no fixed monthly fee or seat pricing, managed infrastructure, automatic scaling, daily backups, dashboard and usage analytics, team collaboration, 99.9% uptime guarantee
EnterpriseContact usBring-your-own-cloud and on-premises deployment, dedicated infrastructure, custom SLA up to 99.95%, up to 24x7 support, 30-minute response SLA, SSO and RBAC, custom integrations, onboarding and training

Is Vectorize worth it?

A credible open source option for agent memory, with a free self-hosted path and metered cloud pricing

Where Vectorize is strong

  • Open source under MIT with no usage limits when self-hosted
  • Per-user memory that persists across sessions and works with any LLM
  • Reported 94.6% score on the LongMemEval benchmark, ahead of the listed alternatives

Where it falls short

  • Cloud pricing is metered per operation, so costs depend on token and call volume
  • No published fixed monthly price or seat-based plan to compare against

How do you get started with Vectorize?

  1. 01Run the Docker command to self-host Hindsight, or start free on Hindsight Cloud
  2. 02Install the skill with npx add-skill vectorize-io/hindsight --skill hindsight-docs to register the MCP memory tools
  3. 03Use the remember, recall and reflect tools, or the Python SDK and REST API, to store and retrieve memories

Vectorize: frequently asked questions

Is Hindsight free to self-host?

Yes. It is MIT licensed with no usage limits, no telemetry and no restrictions, and can run in production on your own infrastructure at no cost.

How does Hindsight Cloud pricing work?

It is pay-as-you-go with no fixed monthly fee and no per-seat pricing. Operations are billed per million tokens or per call, and free credits are provided to start.

What do the individual operations cost?

Retain is $10.00 per million tokens, Recall $0.75 per million tokens, Reflect $0.05 per call, Iris Extract $7.50 per million tokens, Mental Model Retrieve $0.25 per million tokens, Mental Model Refresh $0.05 per call, and storage $0.25 per million tokens per month.

Can I move from self-hosted to Cloud later?

Yes. Vectorize says you can contact them for help migrating memory data from a self-hosted instance to Hindsight Cloud.

Vectorize alternatives

The closest tools in our directory, matched on category, focus areas and audience.

Vectorize compared with its alternatives
ToolBest forFree planStarting priceFocus areas
VectorizeTeams giving LLM agents memory that improves with useYes—Agent Memory, Vector Search
ChromaDevelopers adding vector and full-text search to AI appsYes$250/moVector Search
VersunoUnify all your AI assets in one place.Yes—Agent Memory
Search CandyDevelopers needing citable, versioned document passages for AI agentsNo—Agent Memory
VecRankDevelopers adding vector search and reranking to an appUnknown—Vector Search
AutoflowGraph RAG based conversational knowledge base for TiDB.Unknown—Vector Search
InfinityFlowOffering incredibly fast hybrid search of dense embedding, sparse embedding, tensor and full-text.Unknown—Vector Search
PostgresmlMLOps platform as a PostgreSQL extension for building ML models inside the database.Unknown—Vector Search
ValdInstantly find and analyze vectors with Vald - the customizable AI vector search engine.Unknown—Vector Search
SupermemoryDevelopers adding persistent, cross-model memory to AI agentsYes$19/moAgent Memory, MCP
PineconeDevelopers building vector search, recommendations and RAG applicationsYes$20/moSemantic Search, Vector Search