RunAPI Ideogram V3 MCP server for image generation: create tasks, poll results, and check pricing across 7 model variants from Claude Code, Codex, Cursor, and VS Code.
Two tools with basic schemas and descriptions, but significant gaps in parameter documentation and error handling. get_task has optional parameters with minimal guidance on defaults and dependencies. check_pricing lacks clarity on what 'model' and 'action' mean in context. Neither tool documents output schema structure. No error recovery guidance. Descriptions are present but generic (baseline: 194 chars avg; these are ~100 chars). Parameters lack type constraints (enums, ranges). No validation rules visible. Tool composition is reasonable (single responsibility), but parameter naming could be clearer (e.g., 'action' is ambiguous, does it mean endpoint, operation type, or something else?).
Look up RunAPI pricing for the ideogram-v3 model line.
Fetch the current status and latest result payload for a ideogram-v3 task.
Parameter 'action' is ambiguous in both tools. Does it mean endpoint name, operation type, or something else? Naming should be verb_noun or suffixed with type (e.g., 'endpoint_name', 'action_type'). LLMs will guess wrong.
Output schemas are not documented. LLMs cannot plan downstream calls or extract fields without knowing what get_task and check_pricing return. Document the response structure (fields, types, required vs optional).
Parameter descriptions lack actionable constraints. 'Task id returned when the task was created' does not explain format, length, or example. 'Model slug' does not list valid values or how to discover them. Add enums or format constraints.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 52 | 2026-07-28+ | v2 |
No error handling guidance. If task_id is invalid or model is unsupported, what should the LLM do? Add recovery hints like 'If task not found, verify task_id with the original creation response.'
Optional parameters lack clear defaults. get_task's 'action' is optional and defaults to 'the line's only asynchronous endpoint', but what if there are multiple? Describe the fallback logic explicitly.