Regenerate Workflow

The first generation on a cell is rarely the final one. You’ll iterate — swap the model, tweak the prompt, pass different assets, regenerate. This page covers the patterns and the credit behavior so you know what’s actually happening.

Pattern 1 — update the card, regenerate

If you want to swap models or tweak the prompt, just POST a new generation with the updated data:

curl -X POST "https://api.gen.pro/v1/autocontentengine/$ENGINE_ID/cells/$CELL_ID/generate?agent_id=$AGENT_ID" \
  -H "X-API-Key: $GEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "generation_type": "video_from_text",
    "data": {
      "prompt": "... updated prompt ...",
      "model": "veo_3_1_fast",
      "aspect_ratio": "9:16",
      "duration": 10
    }
  }'

A new generation_id is returned. The previous generation is still there in the history (you can find it via the cell’s generations list), but the cell’s displayed output updates to the new one once it completes.

Pattern 2 — stop and retry

If a generation is running long and you want to bail:

curl -X POST "https://api.gen.pro/v1/generations/789/stop" \
  -H "X-API-Key: $GEN_API_KEY"

Credits for that generation are refunded to your workspace. The generation’s final status becomes stopped. You can then fire a new generation immediately.

Pattern 3 — continue a stopped generation

If you stopped by accident, continue it:

curl -X POST "https://api.gen.pro/v1/generations/789/continue" \
  -H "X-API-Key: $GEN_API_KEY"

Credits are re-charged and the job resumes from pending.

Pattern 4 — mass regenerate

Want to regenerate a whole column across all rows (e.g. you swapped the script prompt)? Use mass_update on the cells + trigger each. No single endpoint regenerates N cells at once — trigger each POST /cells/:cell_id/generate in a tight loop.

Tip

Poll status on the returned generation_id values in parallel. Each generation runs independently on the platform side, so firing 20 at once and polling each completes much faster than doing them sequentially.

Credit behavior summary

Event	Credit behavior
POST /cells/:id/generate	Pre-charge on creation.
Generation completes successfully	Final cost reconciled (refund delta or charge extra).
Generation fails	Full refund.
Generation stopped via /stop	Full refund.
Generation resumed via /continue	Re-charged.
Triggering a new generation on the same cell	Independent charge — previous generation’s credits are not reused.

Check credit balance before a batch

Before kicking off N generations, read the org balance:

curl "https://api.gen.pro/v1/organizations/$ORG_ID" \
  -H "X-API-Key: $GEN_API_KEY"

Response includes available_credit: { generic, aura }. If generic credits are too low for the expected batch cost, top up first — a 422 insufficient_credits_for_job mid-batch is painful.

Seeing all generations for a cell

A cell tracks all its generations, not just the latest. To see the history:

curl "https://api.gen.pro/v1/autocontentengine/$ENGINE_ID/cells/$CELL_ID?agent_id=$AGENT_ID" \
  -H "X-API-Key: $GEN_API_KEY"

Response includes the current user_job_id (the generation currently displayed) and, via the engine’s user_jobs, all historical generations with timestamps and statuses.