Memory layers
- Examples
- Cognee · Mem0 · Zep
- What it does
- Remembers
- Substrate
- Vector + graph stores
- Concurrency model
- Single-tenant memory
- Audit
- Logs
- Reach
- Software-only
● Category map
Three nearby categories. None of them are us.
The clarifying question: Does the system remember context, or does it govern work?
| Memory layers | Agent frameworks | Coding agents | IsoKron | |
|---|---|---|---|---|
| Examples | Cognee · Mem0 · Zep | LangGraph · CrewAI | Cursor · Devin · Claude Code | IsoKron |
| What it does | Remembers | Orchestrates | Edits | Compiles + governs |
| Substrate | Vector + graph stores | Code primitives | Per-agent editor session | 17-entity typed graph + hash chain |
| Concurrency model | Single-tenant memory | Developer-managed | Per-VM / per-worktree | Schema-level claim-fence + lease |
| Audit | Logs | None | None | Per-tenant hash chain + Merkle anchors |
| Reach | Software-only | Software-only | Software-only | Software + hardware-native (HNAO) |
Agent frameworks
- Examples
- LangGraph · CrewAI
- What it does
- Orchestrates
- Substrate
- Code primitives
- Concurrency model
- Developer-managed
- Audit
- None
- Reach
- Software-only
Coding agents
- Examples
- Cursor · Devin · Claude Code
- What it does
- Edits
- Substrate
- Per-agent editor session
- Concurrency model
- Per-VM / per-worktree
- Audit
- None
- Reach
- Software-only
IsoKron
- Examples
- IsoKron
- What it does
- Compiles + governs
- Substrate
- 17-entity typed graph + hash chain
- Concurrency model
- Schema-level claim-fence + lease
- Audit
- Per-tenant hash chain + Merkle anchors
- Reach
- Software + hardware-native (HNAO)
Memory layers — Cognee · Mem0 · Zep
Stores that an agent queries at runtime. We don’t do that. We compile structure before runtime. Zep invalidates old facts as a feature; we hash-chain every fact and let you replay the graph state that made each decision rational at the time. Different layer, different physics.
Agent frameworks — LangGraph
LangGraph is the framework you build an agent in. We’re upstream of where LangGraph sits. We coexist with LangGraph — they run the agent, we decide what the agent should build.
Hosted IDEs / coding agents — Cursor · Devin · Claude Code
These run inside an editor. We compile from natural language to a structured graph that runs against your code via your fleet. Cursor caps parallel mode at 8 via git worktrees; Devin runs managed Devins in isolated VMs. We ship 100-drones-on-a-single-project as the substrate. And code is one of nine entity categories in our graph — workflows, dashboards, manuals, videos, vendor portals, competitor intel, equipment states, regulatory references are the other eight.