| name | predict-engagement-creator |
| description | Focus on sourcing, drafting, and creative selection for: Predict CTR, watch time, comments and shares before production. Use when the Engagement Prediction Creator agent must act on its inputs and produce its defined output. |
predict-engagement-creator
Focus on sourcing, drafting, and creative selection for: Predict CTR, watch time, comments and shares before production.
When to use
- When the Engagement Prediction Creator agent is invoked in the L4 Audience pipeline and its inputs are available.
Inputs / Sources
Historical performance + story metadata. Use the vidiq MCP server's vidiq_video_stats and
vidiq_channel_analytics tools on comparable existing videos as a real calibration baseline for
the prediction, rather than guessing from judgment alone. vidiq is the prioritized subscription;
fall back to nexlev only if vidiq can't answer the lookup.
Output
Draft proposal and creative options for: Predicted engagement metrics
Backend dependency
Writes/reads durable state in PostgreSQL / Apache AGE / PGVector (Railway). BACKEND DEPENDENCY — stubbed until the backbone is wired (spec §2.3/§11.1).
Model
deepseek-direct/deepseek-chat — Cheap high-volume tier (spec §11.3); emit schema-only JSON, restrict to non-sensitive public data (§11.2).
