| @@ -56,10 +56,10 @@ | ||
| 56 | 56 | * |
| 57 | 57 | * The baseline capability gate: a no-network, deterministic |
| 58 | 58 | * `is_supported_for_text_generation()` probe against the AI Client registry |
| 59 | 59 | * (which Core populates from the configured Connectors). This is what plain |
| 60 | - * text-generation features — e.g. comment scoring, which only needs structured | |
| 61 | - * text output — gate on. The agentic assistant needs more; see | |
| 60 | + * text-generation features — the on-demand comment analysis ability, which | |
| 61 | + * only needs structured text output — gate on. The agentic assistant needs more; see | |
| 62 | 62 | * {@see openstation_ai_assistant_provider_configured()}. |
| 63 | 63 | * |
| 64 | 64 | * Credentials are supplied by Core from the configured Connector; no API request |
| 65 | 65 | * is made. |
| @@ -155,11 +155,11 @@ | ||
| 155 | 155 | * |
| 156 | 156 | * - `assistantProviderConfigured` — a provider that supports text generation |
| 157 | 157 | * *and* function calling (what the agentic assistant needs). Gates the Cmd+K |
| 158 | 158 | * assistant, its admin-bar icon, and the "AI assistant" toggle in Features. |
| 159 | - * - `providerConfigured` — the baseline text-generation gate. Comment scoring | |
| 160 | - * (which only needs text output) gates on this; the client uses it for the | |
| 161 | - * "Score new comments with AI" mirror. | |
| 159 | + * - `providerConfigured` — the baseline text-generation gate, for the calls | |
| 160 | + * that need text output but not function calling (the Drafts writing | |
| 161 | + * assistant, the on-demand comment analysis ability). | |
| 162 | 162 | * |
| 163 | 163 | * Provider + model selection is delegated to the Core AI Client, so there is no |
| 164 | 164 | * per-user preference to carry here. |
| 165 | 165 | * |
