Warp Agent Memory
Agent Memory is a fully portable, cross-harness memory system that helps Claude Code, Codex, and the Warp agent remember what works across every session. Now in research preview.
Agent Memory
Warp gives your teams and your entire agent fleet a fast, persistent memory layer that you own.
Fast indexed retrieval
Semantic search on indexed embeddings delivers fast retrieval at an org-wide scale. No slow, agent-driven file traversal required.
Learn how your team works
Memories carry temporal context and relationships, so agents learn how your team works together over time.
Own your long-term memory
Fully portable and auditable memory so you can self-host your memory artifacts on your infrastructure.
Memory in action
Session transcripts are automatically extracted and consolidated so that mistakes, feedback, and decisions are captured in your memory stores. Agents retrieve memories to be more token efficient and effective when they encounter similar tasks in the future.
Self-improving agents
Stop correcting the same mistakes. Agents analyze memory patterns to identify common errors and challenges, then automatically improve agents by updating skills, system prompts, and evals.
Join the Agent Memory Waitlist
Submit the form to join the Agent Memory waitlist.
We love how Warp supports builders through the entire software development cycle and are excited to see them continue to push the frontier.
Marc Manara
Head of Startups at OpenAI
