Developer Open-Sources Bottega Agent Orchestration Tool

A developer open-sourced Bottega, an internal agent orchestration tool used to manage coding agents, after reporting that 100% of their production code has been written by agents for the past 8 months and that the team shipped their 1000th user story with the tool. The tool supports multiple AI models including Codex, OpenCode, Kimi, and DeepSeek, and allows mixing models across pipeline steps such as planning, implementation, and review, reducing sensitivity to top-tier model choice.

Developer Open-Sources Bottega Agent Orchestration Tool In a blog post, the author announces the open-source release of Bottega , an internal agent orchestration tool used to manage coding agents. The post reports that the team shipped their 1000th user story with the tool and that, for the past 8 months , the author reports 100% of their production code has been written by agents. Following Anthropic's new pricing for Claude Code , the author says they added support for Codex and OpenCode to run open-source models such as Kimi and DeepSeek , and to mix models across pipeline steps planning, implementation, review, PR management . The author argues that a robust workflow and harness reduce sensitivity to top-tier model choice and shares the Bottega code and workflow as a minimalist, adaptable orchestration layer. What happened In a blog post, the author open-sourced Bottega , an internal coding agent orchestration tool they built and used in production. The post reports the team shipped their 1000th user story with the tool and that, for the past 8 months , the author reports 100% of their production code has been written by agents. Technical details The post describes Bottega as a minimalist orchestration layer that routes work across models and steps. Per the post, the pipeline separates planning, implementation, code review, and PR management so different models handle each step. The author reports adding multi-provider support after Anthropic changed Claude Code pricing, enabling use of Codex , OpenCode , and open-source models such as Kimi and DeepSeek . The author notes a past bug where Sonnet 3.7 defaulted and output quality remained similar to Opus 4.6 under a strict harness. Editorial analysis Industry observers note that orchestration and process design are the common levers teams use to scale agent-based development. Separating planning, implementation, and review lets teams mix models for cost, latency, or capability tradeoffs without reworking application logic. The shift away from effectively unlimited token budgets makes multi-provider and open-model support a practical resilience strategy. What to watch For practitioners: monitor whether other teams report comparable production usage rates for coding agents, whether open-source orchestration layers gain adoption, and how cost pressure from new pricing drives model mixing or local inference adoption. Watch for tooling that standardizes step-level contracts plan, implement, review across agent stacks. Scoring Rationale Open-sourcing an agent orchestration tool with reported production usage is notable for practitioners building agentized workflows, but it is a single-project release rather than a platform shift. Practice interview problems based on real data 1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with. Try 250 free problems /problems