Overview
Recall is infrastructure for persistent, intelligent agents.
The most powerful agents store memory, track reasoning, and access verifiable data over time. Recall’s blockchain infrastructure powers agents that adapt, evolve, and improve across interactions.
This section is for developers building advanced agents that need structured memory, persistent context, and a reliable infrastructure for long-term learning.
What can you build on Recall?
Every agent is built with four core layers:
- Stateful agents: Build agents that store and retrieve memory across interactions. Learn how agents are structured and where Recall fits in their memory and execution layers.
- Composable memory Use Recall Buckets to structure and manage different types of memory, to help agents learn over time.
- Verifiable data access: Store and retrieve structured or unstructured data used in training, inference, or reasoning workflows.
- Protocol-integrated agents: For advanced use, plug into Recall's on-chain protocol for tamper-proof storage, data lineage, and persistent context flows.
Why develop your agent on Recall?
- Open storage: Anyone can create and query memory buckets with public or permissioned access.
- Verifiable context: Blake3 hashes ensure memory is cryptographically consistent and tamper-proof.
- Composability: Compatible with any agent framework, across both web2 and web3 stacks.
Get started
Build stateful agents
Discover how to create intelligent agents with Recall's stateful agent framework.
Data sources & Buckets
Learn how to structure and organize your agent's memory using Recall Buckets.
Protocol architecture
Dive into Recall's protocol and the underlying blockchain infrastructure.
Developer tools
Explore SDKs, CLIs, and development tools for building on Recall.