SenteLabs releases OpenExecutive v0 on GitHub with eight Claude-backed specialist agents
OpenExecutive provides production-grade agent orchestration and memory for executive tasks using current Anthropic models and local vector stores. The release demonstrates immediate developer response to displacement by shipping verifiable, auditable AI tooling under Apache 2.0. Architecture choices prioritize cache efficiency and single-instance scheduling over horizontal scaling.
Developers published the full repository after reported firings tied to AI substitution. The system routes every user message through an Executive Orchestrator that calls eight parallel agents (CSO, CFO, CHRO, GC, COO, CMO, CPO, Board) each performing two-layer retrieval from builtin MBA Markdown and company_docs collections before synthesizing a single voice response. Episodic memory extraction runs on claude-haiku-4-5 after every turn and a single-instance scheduler claims due actions via RETURNING to surface follow-ups.
Anthropic: claude-sonnet-4-6 usage in the repo exceeds 10k calls per week within 90 days of release.
Sources (2)
- [1]Primary Source(https://github.com/SenteLabsAI/OpenExecutive)
- [2]Supporting Source(https://docs.anthropic.com/en/docs/build-with-claude/prompt-caching)