Kooma — Bambara AI

Each tool targets a clearly different capability: listing voices, synthesizing speech, and translating text. Descriptions explicitly state the dependency (list_voices must precede text_to_speech), which removes any ordering ambiguity. An agent can select the right tool without guesswork.

All names share a consistent kooma_ prefix and snake_case formatting. Verb/noun ordering is slightly mixed (kooma_list_voices is verb_noun, while kooma_text_to_speech and kooma_translate are noun/verb forms), but the pattern remains readable and predictable.

Three tools is on the lean side but each earns its place: a lookup, a synthesis action, and a translation action. For a narrowly scoped Bambara language service this surface is reasonable, though a couple more operations would round it out.

The TTS workflow is self-contained (list voices, then synthesize) and translation covers 14 languages, but there are notable gaps: no speech-to-text/ASR, and no tool to enumerate supported language pairs or check character limits for synthesis. Agents can work around these using the descriptions, but the surface stops short of full lifecycle coverage.