Engineering blog
Deep dives on atom-based memory, MCP, and reproducible benchmarks—written for builders shipping agents to production.
Common Knowledge: a shared memory library for every team
Ship platform-wide facts once—product docs, runbooks, conventions—and let every workspace merge them into search, recall, and wake-up by default. Read-only for customers; writable only by platform operators.
Correcting memory atoms: versioned updates instead of duplicate saves
PATCH /atoms/{atomId}, MCP memory_correct_atom, optimistic concurrency with version, per-library auto-correct policy, and archived revisions in R2.
Engra is not Engram: disambiguating arXiv:2606.09900
Engra (engra.ai) is a production agent memory SaaS by Artalaxies LLC. The open-source Engram engine in arXiv:2606.09900 is a separate research project—same problem space, different model and benchmarks.
Why structured atoms beat fixed-size chunk RAG
On ENGRA-KB-v2 (550 docs, 400 queries), structured atoms reach 75.3% Recall@10 vs 54.8% for chunk RAG—how Library → Scope → Topic → Atom changes retrieval.
Choosing an Engra plan: Starter, Growth, Pro, and Enterprise
Transparent USD pricing—when to stay on Starter, move to Growth for Memory-heavy teams, scale to Pro, or talk to us for Enterprise custom quotas.
The L0–L3 wake-up stack: how agents load the right memory depth
L0 identity, L1 narrative, L2 topic recall, L3 deep search—when to call wake_up vs recall vs search, and why flat RAG context windows waste tokens.
Library, Scope, Topic, Atom: a memory model that scales with your agent fleet
Four layers answer whose memory, which domain, which subject, and which fact—how to model multi-tenant agents without mixing customer contexts.
Memory MCP: five-minute Cursor setup
Streamable HTTP MCP tools for libraries, search, recall, wake-up, and atom writes—plus SKILL.md workflows for production agents.
Memory quotas explained: upload, storage, retrieve, and preview
How Engra maps R2 storage economics to monthly upload and atom storage caps, plus retrieve and source-preview limits that align with wake_up, search, and recall.
Gateway logical routes: one HTTP surface, many agent backends
How logic.json routes requests by path, method, and headers to workflows, memory, or external APIs—without redeploying your gateway for every new agent.
Gateway production deployment: staging keys, routes, and rollout
A practical checklist—separate gateways per environment, test logic.json in staging, rotate API keys, and wire Memory bindings before you point customer traffic at Engra.
Agent Skills in production: wake, search, save as a repeatable workflow
SKILL.md patterns for Cursor and Claude Code agents—session bootstrap, topic recall, atom writes, and guardrails that keep memory consistent across engineers.
Reproducing ENGRA-KB-v2 benchmarks on your machine
Download the JSON suite, run the same embedder and query set, and compare structured atoms vs chunk RAG before you migrate production memory.