RunAPI Qwen Image MCP server for image generation: create tasks, poll results, and check pricing across 3 model variants from Claude Code, Codex, Cursor, and VS Code.
Server has 4 tools with mixed quality. Two static tools (get_task, check_pricing) have descriptions and basic schemas, but two dynamically-generated tools (dynamic_asynchronous_endpoints, dynamic_synchronous_endpoints) lack visible input schemas and detailed descriptions. Parameter descriptions are sparse or missing. No error handling guidance visible. Tool names lack action verbs (e.g., 'dynamic_asynchronous_endpoints' is vague). Output schemas not documented. This is below-median for production use.
Look up RunAPI pricing for the qwen-image model line.
Dynamically generated tools for asynchronous endpoints from contract.json. Each endpoint creates a task on RunAPI and returns a task id, status, and output URLs.
Dynamically generated tools for synchronous endpoints from contract.json. Each endpoint runs a synchronous operation on RunAPI and returns the operation result.
Fetch the current status and latest result payload for a qwen-image task.
Two dynamically-generated tools lack visible input schemas. Tool definitions are inferred from contract.json but no explicit schema registration shown in source code.
Tool names lack action verbs. 'dynamic_asynchronous_endpoints' and 'dynamic_synchronous_endpoints' do not start with verbs (create_, run_, execute_). LLMs cannot infer intent from these names alone.
Output schemas not documented. No visible documentation of what get_task, check_pricing, or the dynamic tools return. LLMs cannot plan downstream calls or extract required fields without knowing response structure.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 45 | 2026-07-28+ | v2 |
Error handling lacks recovery guidance. Code returns jsonText({error: ...}) but does not categorize errors as retryable/user-fixable/fatal or suggest next steps. E.g., 'Unsupported RunAPI service/action/model combination' does not guide the LLM to call search or list tools.
Parameter descriptions incomplete. 'action' parameter in get_task and check_pricing has minimal description. No constraints on format, length, or valid values documented.