How to Use Higgsfield: A Beginner's Guide (2026)

TL;DR — Pick Image or Video at the top, pick a model, write a prompt, and press Generate to finish your first run. Everything else on the screen is a specialised studio that you don't need yet. The main trap for beginners is buying the $19 Starter plan for a model it doesn't include.

The home screen opens on eight studios and a node board, and none of them is the one that makes your first video. Your first generation uses only three controls right on the main page. We opened a free account on 24 August 2026, and the screenshots below show what you'll see on that account.

What you need before you start

You'll need an account, a web browser, and a clear idea of what you want to make. Higgsfield calls itself "an AI-native creative suite" that "combines proprietary tools for video production, character consistency, and ad creation with access to a growing library of AI models under one subscription". It gives you access to outside models like Seedance, Kling, Veo, Sora 2, and Nano Banana alongside its own tools. You won't need any video editing, design, or machine learning background to make your first clip.

Before you spend any money, you'll want to check the plan tiers closely so you don't pick the wrong one. The $19 Starter plan card carries the heading "NO ACCESS TO SEEDANCE 2.5 — Available from Plus plan", and it lists both Seedance 2.5 and Seedance 2.0 as "No access".

Here's how Higgsfield splits its core access and limits:

Free plan: listed as "Limited use" and runs 1 job at a time.
Starter plan ($19): runs up to 2 videos or 4 images at once, but only includes Seedance 2.0 Fast and 2.0 Mini.
Plus plan ($59): opens Seedance 2.5 and full Seedance 2.0, with up to 6 jobs at once.
Ultra plan: opens the highest limits and lets you run up to 8 jobs at once.

Higgsfield's pricing page with the Starter plan card reading NO ACCESS TO SEEDANCE 2.5, next to the Plus and Ultra cards that both show full access Higgsfield's own pricing page, 24 August 2026. On the $19 Starter plan, Seedance 2.5 and Seedance 2.0 both read "No access".

Step 1. Pick Image or Video at the top

Higgsfield sums up the whole creation flow in one simple sentence: "Choose Image or Video from the top navigation, select a model or feature, enter a prompt or upload a reference, and click Generate."

Click Video in the top navigation bar to open the main creation screen. The page turns into a clean prompt box that reads "Describe any visual idea. We will generate a video."

Directly below the prompt box, you'll see a row of basic controls:

A model chip showing the active generator engine.
A duration chip that defaults to 8s.
An aspect-ratio picker set to Auto.
An audio toggle switched to On by default.
A bright Generate button on the right edge.

Higgsfield's AI video generator with the prompt box, the model chip set to Seedance 2.0, the duration chip, the aspect control and the Generate button Higgsfield's video generator on our own free account, 24 August 2026. Prompt box, model, duration, shape, audio, Generate — that's the whole first run.

Step 2. Choose a model or a feature

Click the model chip next to the prompt box to see what your plan lets you run. On our free test account, the video generator loaded with Seedance 2.0 as the default choice.

You can stick with that default or click through the list to pick outside tools like Kling, Veo, Sora 2, or Nano Banana. If you're on the Starter plan, standard Seedance 2.0 and Seedance 2.5 stay locked until you upgrade to Plus or Ultra.

Step 3. Write the prompt

Click inside the text box and write what you want to see. Higgsfield explains prompts this way: "A prompt is a text description of what you want the AI to create. The more specific, the better."

If you aren't sure how to start, look at Higgsfield's own official prompt examples:

For video: "A close-up of a dancer mid-spin on a rooftop at golden hour, slow motion, cinematic lighting, shallow depth of field."
For images: "Fashion editorial portrait, studio lighting, high detail, magazine photography."

You can paste the same prompt on Seadanse to see how different engines handle the exact same scene description. State the main subject, the lighting, the background, and the framing clearly so the model knows what to create.

Step 4. Add a reference image if the model takes one

Look beside the prompt box for an image upload button if your chosen model supports reference files. A reference image is a picture the model copies the look, style, or subject from during generation.

You won't need an image for a basic text run, so you can skip this step entirely if you're only turning words into video. If you do upload a file for an in-house tool, save it as a JPG, PNG, or WebP picture that stays within the size limit so it doesn't throw an error.

Step 5. Set the duration, the shape and the resolution

The settings chips sit directly under the prompt box. On our free account, the duration chip read 8s and the aspect-ratio control read Auto.

The aspect options offered were 16:9, 9:16, 4:3, 3:4, 1:1, and 21:9. The audio toggle was set to On by default, and you can turn it off if you don't want sound on your clip.

Step 6. Press Generate and find it in Assets

Click the Generate button on the right side of the control bar to start your generation job. That sends your prompt to the processing queue right away.

Most generations finish within 30 seconds to 10 minutes. Image generations are usually faster, while videos take longer, especially if you pick higher output resolutions.

Once the job finishes, your result shows up directly inside the section where you generated it and also on your dedicated content page. On a fresh account, that page sits empty with a note reading "Your generations will appear here". Once your video finishes, you'll find it there ready to play, inspect, or download.

Higgsfield's Assets page on a new account, empty, with the message that your generations will appear here Assets on a new Higgsfield account, 24 August 2026. Everything you make lands here.

The studios you can ignore on day one

Higgsfield packs several specialized workspaces onto its home screen. You won't need any of them to make your first basic images or video clips.

Soul and Soul ID

Soul is Higgsfield's own image model family built for film-style, fashion, and culture-native visuals, and Soul ID lets you train a character from photos so the same face appears across every generation.

Cinema Studio

Cinema Studio gives you tools for camera control, multi-shot scenes, and physics-aware motion on its own dedicated screen.

Marketing Studio

Marketing Studio turns a product URL or image into a finished video ad in UGC, unboxing, tutorial, or TV-spot formats with audio and lip sync.

Canvas

Canvas is a node-based board where you chain models in sequence, collaborate in real time, and save these chains as reusable templates.

Supercomputer

Supercomputer takes a plain-language project description and assigns an agent to plan the project, run the models, and deliver finished files.

Higgsfield Audio

Higgsfield Audio handles voiceover generation, voice cloning, and audio translation.

Settings that actually change your result

Beyond the prompt words, a few basic switches change how your generation runs and what the file looks like:

Aspect ratio: picking 9:16 instead of 16:9 changes what the model fits in frame and shifts the camera perspective.
Queue mode: credit-based runs use the priority queue for top speed, while Unlimited runs use the standard queue.
Credit mode: turning off Unlimited and switching to Credit Mode bypasses standard wait times and returns to maximum speed instantly.
Jobs at once: your plan sets your jobs at once to 1, 2, 6, or 8, though you can buy a Concurrency Boost to raise it.
Subscription cycles: monthly subscription credits do not roll over and expire at the end of each credit cycle.

If a run ever gets stuck, you won't lose what you spent on it. Credits for failed generations are returned automatically, and you can check Manage Account and Usage for a "Refunded" entry.

Four things beginners get wrong

Buying the Starter plan for Seedance 2.5: the $19 Starter plan marks both Seedance 2.5 and standard Seedance 2.0 as "No access", offering only Fast and Mini variants.
Letting unused credits expire: monthly subscription credits don't carry over into your next billing period, so unspent credits vanish at cycle end.
Uploading bad file types to in-house tools: using formats other than JPG, PNG, or WebP causes in-house features to fail.
Waiting on frozen queue jobs: most clips take under 10 minutes, so if a tile hangs, hover over it, click Cancel in the top-right corner, and retry.

Try the same idea somewhere else

On the Seedance 2.5 AI Video Maker at Seadanse, every paid plan opens the exact same eight models: Seedance 2.5, Seedance 2.0, Veo 3.1, Wan 3.0, Hailuo 3, Grok Imagine 1.5, Nano Banana 2, and GPT Image 2. No tier withholds a model.

The Seadanse composer with the prompt box, model picker, duration and resolution controls above the Generate button The composer on Seadanse, 24 August 2026.

You can also run Seedance 2.5 without a subscription by buying a one-time pack of 500 credits for $9.9 without any recurring monthly fee. Check what each plan includes before choosing how you want to create.

The Seadanse composer keeps every setting on one screen:

A prompt box that holds up to 10,000 characters.
Six fixed clip lengths: 5, 10, 15, 20, 25, or 30 seconds, with no extension.
Six aspect ratios: 16:9, 9:16, 1:1, 4:3, 3:4, and 21:9.
Three resolution choices: 480p, 720p, or 1080p, with no 4K on Seedance 2.5.
Audio turned on by default.

New accounts get free starter credits with no card, so you can paste your prompt and try one there.

Higgsfield Not Working? 7 Fixes That Actually Work — what to do when your first generation hangs.

7 Best Higgsfield Alternatives in 2026 — what each rival charges, and what happens to credits you don't spend.

How to Cancel a Higgsfield Subscription — the click path, and what happens to credits you already paid for.