Bring the agent you already pay for.
AI workers run on your own Claude Code subscription and licence. We sell the harness and the memory—so there is never an inference bill from Metantel.
Already subscribed? Your plan continues unchanged. Manage it in your console →Trial versions of our Harnesses.
- ✓ Flow Console trial on your machine
- ✓ Artel QA Lite (local via npm)
- ✓ Runs on your own Claude Code subscription
- ✓ Zero inference markup from Metantel
For individual developers cutting token bloat.
- ✓ Full Flow Console desktop orchestrator
- ✓ Personal EDIP agent-paced memory (−47% session cost)
- ✓ Artel QA source-aware verification
- ✓ Graduate sessions into persistent AI Workers
Shared memory, intent collaboration and the org ladder.
- ✓ Everything in Personal, across devices
- ✓ HPC MOESI function-level cache coherence across laptops
- ✓ Cross-machine agent-to-agent messaging & pre-edit conflict resolution
- ✓ 4-layer Team Memory & Org Ladder access scoping
- ✓ Artel Workspace & EDIP Workflows (Beta)
Self-hosted memory, custom workflows & Agentify MCPs.
- ✓ Self-hosted / VPC deployment inside your perimeter
- ✓ Custom EDIP Workflows & multi-modal connectors
- ✓ Agentify: governed web-app hosted MCP APIs
- ✓ Any model provider or internal reasoning models
No inference bill
AI workers run on the customer’s own Claude Code subscription and licence. We sell the harness and the memory—so your token savings go straight back into your budget.
Priced where buyers are
Comparable enterprise agent infrastructure (such as Microsoft Agent 365) prices at $15–$30 per user per month. With Metantel, long-session token savings (−47%) often exceed the seat price itself.
Land a squad, grow with the team
Start with 2–4 developers and zero backend setup. Because what one developer’s agent learns flows into shared memory, the platform gets more valuable with every teammate added.