Mycel launches an open-core AI platform designed for service businesses to draft client deliverables, automate back-office workflows, and learn from human review inside secure sandboxes.
Mycel is an open-core AI automation platform built for service businesses, consultancies, and agencies seeking to automate recurring client deliverables without sacrificing quality control. By uploading a single past deliverable—such as an SEO audit, client report, contract review, or monthly financial close—Mycel learns the firm's structure and standards to automatically generate subsequent drafts across accounts. Each task executes in an ephemeral, credential-free sandbox to protect client data privacy, and all outputs require human approval before being sent. The platform tracks operator edits to improve future drafts over time and offers white-labeled client portals. Hosted plans start at $299 per month, while the core engine is released under the Apache-2.0 license for free self-hosting using custom model API keys.
Service agencies do not need generic chatbots; they need review-gated assembly lines that capture firm-specific taste and improve through iterative human correction.
• Human-in-the-Loop Quality Control: Requiring explicit approval before anything ships and measuring rewrite percentages over time grounds the AI's utility in real accuracy metrics.
• Disposable Sandbox Architecture: Isolating execution environments without storing persistent client credentials directly mitigates client data leakage risks.
• Open-Core Distribution: Releasing an Apache-2.0 self-hostable engine allows privacy-conscious agencies to keep client data on-premises while providing a clear monetization ramp via managed hosting.
• Direct Agency Economics: With white-label client portal support and pricing starting at $299/month, the platform positions itself as an alternative to the recurring overhead and training required for junior account managers.
DISCOVERED
1h ago
2026-09-20
PUBLISHED
7h ago
2026-09-20
RELEVANCE
AUTHOR
ISLAM HACHIMI