AI-powered website builder that generates a professional site in under a minute and includes tools for content, SEO, and analytics.
52M visits/mo+0.7%#4 Coding & Development
Polign Recall
Typed memory layer for AI agents with correction rules and queryable history
$12k/yrStarting price
About Polign Recall
Polign Recall provides a memory infrastructure layer for AI agents that stores typed facts with explicit correction rules and maintains a queryable history of changes. The system operates on top of polign_db, which keeps all indexes in the user’s own object storage, treating serving nodes as replaceable for scaling flexibility. Data remains under the user’s control at all times, addressing sovereignty concerns for organizations. It supports multiple agents or processes reading and writing the same memory simultaneously within a user-owned namespace. The memory layer is designed for scenarios where facts change over time and requires no hosted service signup, instead focusing on self-hosted deployment. Polign Recall is open source under Apache 2.0 and runs on user-provided compute infrastructure.
Key features
- Typed memory with explicit correction rules
- Queryable history of all previous values
- Multi-agent read/write access
- Self-hosted deployment on user infrastructure
- Open source under Apache 2.0
- Data stored in user-owned object storage
- Encryption with user-held keys
- Layered caching for repeat queries
Use cases
- Long-lived memory about users or accounts that changes over time
- Shared memory across multiple agents or processes
- Self-hosted or private cloud deployments with data sovereignty requirements
Pros
- Typed memory with explicit correction rules instead of model-based judgments
- Queryable history of all previous values and beliefs
- Data stored in user-owned object storage with full control
- Supports multiple agents across different processes and machines
- Open source under Apache 2.0 license
Cons
- No production deployments currently exist
- Requires self-hosting on user's own infrastructure
- Not suitable for hosted service signups
- Does not automatically extract facts from conversations
Frequently asked questions about Polign Recall
What does Polign Recall do for AI agents?
Polign Recall provides a memory infrastructure layer that stores typed facts with explicit correction rules and maintains a queryable history of changes for AI agents. It allows agents to remember and update information while preserving a record of past beliefs.
Who is Polign Recall designed for?
Polign Recall is designed for teams running AI agents in or near production who need long-lived, changeable memory, shared access across multiple agents, and data sovereignty. It suits organizations with self-hosted or private cloud requirements.
How does Polign Recall handle data storage and ownership?
All data and indexes remain in the user’s own object storage (e.g., S3, GCS, MinIO) under their control. The system treats serving nodes as replaceable, ensuring data stays in the user’s account and is never hosted by a third party.
Can multiple agents or processes access the same memory simultaneously?
Yes, Polign Recall supports multiple agents or processes reading and writing the same memory within a user-owned namespace, enabling shared access across different machines and processes.
Is Polign Recall a hosted service or self-hosted?
Polign Recall is designed for self-hosted deployment and does not require a hosted service signup. Users can run it on their own infrastructure, including premises, cloud, or Kubernetes environments.
What are the pricing options for Polign Recall?
Self-hosting is free to use indefinitely. For production support, there is a flat annual fee with no metered charges. The Design Partner Program offers seats at no charge for two months, followed by a paid plan.
