HStart for $8/month
Guide · FLUX · Wavespeed

Setting up FLUX with a Wavespeed API key

FLUX is one of the strongest open image model families available through the Wavespeed gateway. Hatch AI Studio is BYOK, so you pay Wavespeed directly for generations with no model markup. This guide takes you from creating a Wavespeed API key to your first high-quality FLUX render — and then into a Canvas workflow.

You need one thing before you start: a Wavespeed account with a funded API key. Hatch never charges per generation — provider usage is billed on your own key.

  1. Step 1

    Create a Wavespeed account and API key

    Sign up at wavespeed.ai, open the API keys section of your dashboard, and generate a new key. Copy it once — Wavespeed shows the full key only at creation time. Add a small amount of credit so your first FLUX call isn't rejected for insufficient balance.

  2. Step 2

    Add the key to Hatch AI Studio

    In Hatch, open Settings → API keys and paste the Wavespeed key. Keys are stored server-side against your account, never in browser localStorage, and are only used to call Wavespeed on your behalf. Save, then reload the studio so the model list refreshes.

  3. Step 3

    Choose the right FLUX model

    Open the Image studio and pick a FLUX model from the picker. FLUX Dev is the balanced default for general prompts, FLUX Schnell is the fastest and cheapest for drafts and iteration, and FLUX Pro / Ultra give the highest fidelity for finals. The per-run cost estimate updates as you switch models.

  4. Step 4

    Set resolution and reference images

    Pick an aspect ratio and resolution before generating — higher resolutions cost more per run. FLUX editing and redux variants accept reference images: attach them from your uploads or previous generations with the media picker, or call them inline in the prompt with @.

  5. Step 5

    Run your first generation

    Write a specific prompt (subject, style, lighting, lens, mood), confirm the estimated cost, and hit Generate. The result lands in History with the prompt, model, and parameters preserved so you can re-run or fork it later.

  6. Step 6

    Chain FLUX into a Canvas workflow

    In Canvas, add an Image Gen node, select your FLUX model, and wire its output into video, lip sync, or upscale nodes. Run a single node while iterating, then Run All once the graph is right.

Troubleshooting a failed FLUX run

  • 401 / unauthorized: the key was pasted with whitespace or was revoked. Re-copy it from the Wavespeed dashboard and save again in Settings.
  • Insufficient balance: Wavespeed rejects calls when credit runs out. Top up, then re-run from History.
  • Long-running job: high-resolution FLUX Pro renders can take a while. Hatch keeps polling and recovers the result — leave the tab open or check History.
  • Reference image rejected: not every FLUX variant accepts references. Switch to a FLUX edit/redux model when you need image-to-image.

Ready to run your first FLUX render?

Add your Wavespeed key once and every studio — image, video, lip sync, and Canvas — unlocks at the same time.

Open Hatch AI Studio