What I Learned Building Co-op Translator MCP
I recently added a Model Context Protocol server to Co-op Translator. Co-op Translator started as a CLI-first tool for translating Markdown files, Jupyter notebooks, and text inside images. The CLI is still the most natural interface when a person wants to translate a repository from a terminal. But agents see tools differently. When an agent such as Codex or Claude Code works with a project, it…
Co-op Translator added a Model Context Protocol (MCP) server, which began as a CLI tool for translating various file types. However, when considering agent-assisted translation, the need for safer defaults and clearer boundaries between human and agent interactions became apparent. Separating responsibilities into smaller, model-controlled tools improved the API design and made it easier to understand and recover from translation processes.
Agent-assisted translation allows Co-op Translator to prepare work, while the host agent handles actual translation, making the most of each party's strengths. Image translation, however, requires provider-backed vision capabilities, so it remains separate from the agent-assisted workflow. The main takeaway is that agent-facing APIs should prioritize composability, predictability, and safe defaults, rather than simply mimicking the convenience of human-facing CLIs.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
