Krea Client Workspace Setup (Onboarding)

Krea Client Workspace Setup (Onboarding)

1. 01

   Create the client workspace account and confirm the plan tier before any generation work starts

   Free tier caps at 100 units per day with LoRA training limited to 50 images and upscaling capped at 2K, so any retainer client needing print-ready output or brand model training must be provisioned on Pro at $21/mo (annual billing) or Max at $105/mo. Commercial licensing is available across all tiers, so the tier decision is driven by unit volume and upscaling ceiling, not by licensing.

2. 02

   Assign the frontier model set the client's deliverables will draw from

   Krea aggregates Krea 2, Nano Banana Pro, Seedance 2.5, Veo 3.1, FLUX, and MiniMax in one workspace. Pin the specific image and video models per client so two agencies running the same account do not produce mismatched visual styles across a campaign.

3. 03

   Train the client LoRA model on brand assets

   Upload the approved brand asset set (50 images on the standard LoRA training allowance) and run training before generating any client-facing asset. This locks the brand style into the model so downstream generations do not drift between operators on the account.

4. 04

   Build the delivery pipeline in the Node Editor

   Recreate the client's repeatable workflow as a node graph covering generation, Krea Enhancer upscaling, and export. Template workflows exist for architectural renders, commercial mockups, and explainer videos; clone the closest match and modify rather than building from an empty canvas.

5. 05

   Set the Krea Enhancer upscaling ceiling per deliverable type

   Pro supports upscaling to 22K and the Enhancer is documented to 21K. Match the ceiling to the output channel: social assets stay low, print-ready 4K outputs for the client library get the higher pass. Record the setting in the client's pipeline notes so delivery staff do not guess.

6. 06

   Document prompt templates and the export workflow for handoff

   Capture the prompt strings, model selections, LoRA reference, and node graph as a written runbook. This is the artifact that lets a second operator on the agency team reproduce the client's output without re-training or re-testing models.

7. 07

   Run a pilot batch and reconcile unit consumption against the plan allowance

   Pro includes 20,000 units per month. Generate the pilot set, then check consumption against that figure before committing the client to a monthly deliverable volume. If the pilot exceeds the allowance, the retainer scope or the tier needs to change before the first invoice.