Interfaces

Console

Use the Skytells Console to experiment with AI models through a fully interactive playground — no coding experience required.

What Is the Console?

The Skytells Console is a fully interactive, browser-based playground that communicates directly with Skytells APIs. It's built for everyone — from developers prototyping their next product to designers, researchers, content creators, and complete beginners who have never written a line of code.

Think of it as a control panel for every AI model on Skytells. You pick a model, fill in a form, click Run, and see the result instantly — no terminal, no IDE, no deployment. The Console handles authentication, request formatting, parameter validation, and output rendering so you can focus entirely on what you're building.

Why Use the Console?

  • Zero coding required — every parameter is exposed as a visual form field (text inputs, dropdowns, sliders, toggles). Just fill in the blanks and hit Run.
  • Instant visual previews — generated images render inline, audio plays back directly, video streams in the browser, and text outputs are formatted and syntax-highlighted. You see the result exactly as your users would.
  • Built-in code generation — once you've crafted the perfect input in the form, switch to the Node.js, Python, or HTTP tab to grab a production-ready code snippet. The Console writes the code for you.
  • Full schema documentation — every model's input and output schema is browsable right in the Console. See every parameter, its type, description, default value, and constraints without leaving the page.
  • Transparent pricing — costs are shown before you run. You always know what a request will cost.
  • Experiment freely — try different prompts, tweak parameters, compare outputs, and iterate in real time. The Console is designed for rapid experimentation.

The Console is available at skytells.ai/dashboard/console. Open any model from the model catalog and you'll land directly in its Console page.


Who Is the Console For?

The Console is designed to be approachable for every experience level:

AudienceHow They Use the Console
Complete beginnersFill in a text prompt, click Run, and see AI-generated images, text, or audio — no setup needed
Designers & creativesRapid visual prototyping — generate images, iterate on prompts, compare aspect ratios and styles
DevelopersPrototype API calls visually, then copy the auto-generated code into their apps
Product teamsEvaluate models before committing to integration — test quality, latency, and cost
ResearchersExplore model capabilities, test edge cases, and inspect raw JSON responses

Prerequisites

Before using the Console, make sure you have:

An API Key

You need an active API key. If you don't have one, follow the Generating an API Key guide to create one from the Dashboard.

A Minimum Balance of $2

Your account must have at least $2 in credit to run models through the Console. You can add credit from the Billing page.

If your balance drops below $2, the Run button will be disabled. Top up your account to continue using the Console.


Console Layout Overview

When you open a model in the Console (for example, Skytells / TrueFusion), the page is organized into several distinct areas. Here's a complete walkthrough:

At the very top of the page, a breadcrumb trail shows your location:

Home > Dashboard > Console > [Model Name]

Click any segment to navigate back to that level.

Model Header

Directly below, a prominent header displays:

  • Model avatar and name — e.g., Skytells / TrueFusion
  • Model type badge — appears in the top-right corner (e.g., IMAGE, TEXT, AUDIO, VIDEO)
  • Back arrow — click to return to the Console model list

Model Info Badges

A row of badges provides critical details at a glance:

BadgeWhat It Tells You
✓ OfficialThis model is built and maintained by Skytells
ⓘ AboutLinks to the model's documentation and details page
✓ OperationalThe model is online and accepting requests. If it shows "Degraded" or "Down", expect issues
💳 $0.03 / imageThe cost per unit of output — varies by model (per image, per token, per second, etc.)
🖥 H100The GPU hardware the model runs on (e.g., H100, A100, T4)
🌐 PublicVisibility — "Public" means anyone can use it; "Private" means restricted to your organization

Top Controls

Below the badges, three controls sit side by side:

ControlDescription
Model dropdownSwitch to a different model without leaving the Console (e.g., switch from TrueFusion to another model)
API Key fieldYour API key is pre-filled and masked with dots. Click the 👁 eye icon to reveal it. This is the key used for all Console requests
▶ Run buttonExecutes the model with your current inputs. Results appear in the Output panel within seconds

Tab Navigation

Each model's Console page has four main tabs. These tabs let you move between running the model, inspecting its schema, browsing examples, and reading documentation — all in one place.

TabIconWhat It Shows
PlaygroundThe interactive runner — this is where you input parameters and execute the model
API📋Full schema browser — input/output parameter definitions, types, constraints, and defaults
ExamplesCurated input/output pairs — see what good prompts look like and what the model produces
README📖Model documentation — usage notes, capabilities, limitations, and changelog

The Playground — In Detail

The Playground tab is the heart of the Console. It's a fully interactive environment split into two panels: Input on the left and Output on the right.

Input Panel (Left Side)

The input panel is where you construct your request. It offers five different views, and you can switch between them at any time using the tabs at the top of the panel:

The Form view renders every model parameter as a visual form field. This is the default view and the best starting point for everyone — especially beginners.

  • Text fields for string parameters (e.g., prompt)
  • Dropdowns for enum parameters (e.g., aspect_ratio with options like 1:1, 16:9, 4:3)
  • Sliders for numeric parameters (e.g., number_of_images with min/max constraints)
  • Toggles for boolean parameters (e.g., prompt_optimizer on/off)
  • Required fields are clearly marked with a red asterisk
  • Descriptions appear below each field explaining what it does
  • Advanced Options — click to expand and reveal additional, optional parameters

You don't need to know JSON, APIs, or programming. Just fill in the fields and click Run.

The JSON view shows the raw request payload as editable JSON. Useful for developers who want precise control over the input structure, or for pasting in a payload from another source.

The Node.js view shows an auto-generated code snippet that reproduces your current form inputs as a Node.js/TypeScript API call using the Skytells SDK. Copy this directly into your project.

The Python view shows the equivalent Python code snippet using the Skytells Python client. Same inputs, ready to paste into your Python project.

The HTTP view shows the raw cURL command for the request — including headers, URL, and JSON body. Useful for debugging, sharing with teammates, or using from any language.

Example: TrueFusion Image Model Parameters

When using the TrueFusion image generation model, the Form view shows:

ParameterTypeDescriptionDefaultConstraints
promptstringText prompt for image generationrequired
aspect_ratiodropdownImage aspect ratio1:11:1, 16:9, 4:3, 3:2, 2:3, 3:4, 9:16, 21:9
number_of_imagessliderHow many images to generate1min: 1, max: 9
prompt_optimizertoggleAutomatically enhance your prompt for better resultstrue

The prompt field is the primary input and always visible. Additional parameters like aspect ratio, image count, and prompt optimizer are under Advanced Options — click to expand them.

Output Panel (Right Side)

After clicking ▶ Run, the model processes your input and results appear in the output panel. The output panel has three tabs that let you inspect the result in different ways:

TabWhat It Shows
👁 PreviewA rich, visual rendering of the model's output — the format adapts to the output type (see below)
JSONThe raw JSON response from the API — every field, metadata, and URL
>_ LogsExecution logs including timing, request ID, and debug information

Before a model has been run, the Preview tab shows a placeholder: "Preview will appear here after running the model."


Output Preview Types

One of the most powerful features of the Console is its intelligent output preview. The Preview tab automatically detects the output type and renders it with the appropriate viewer — so you always see results in their native format, not as raw JSON blobs.

Output TypeWhat the Preview Shows
ImagesGenerated images rendered inline at full resolution. Multiple images display in a grid. Click to enlarge.
TextFormatted text output with syntax highlighting for code, markdown rendering for prose
AudioBuilt-in audio player with play/pause, seek, and download controls
VideoInline video player with playback controls, frame-by-frame seeking
Structured dataFormatted tables or JSON trees for structured outputs (e.g., classification results, extracted entities)
FilesDownload links for generated files (PDFs, CSVs, etc.)

This means a designer generating images sees a visual gallery, a developer working with text sees formatted output, and a researcher analyzing structured data sees clean tables — all without touching any code.


Browsing the API Schema

Switch to the API tab to explore a model's full schema. This is like built-in API documentation for every model on Skytells.

The left sidebar in the API tab has five sections:

SectionWhat It Contains
PlaygroundQuick link back to the interactive runner
SchemaDetailed input/output parameter definitions — the technical spec for the model
API ReferenceEndpoint URLs, authentication headers, and request format for this specific model
ExamplesCurated input/output pairs showing real usage
READMEFull model documentation

Schema Browser

The Schema view has two sub-tabs:

  • Input schema — every accepted parameter displayed as a table with columns for Name, Type, Description, Default, and Constraints. Toggle between Table and JSON views.
  • Output schema — the structure of the response payload so you know exactly what to expect

This is especially useful for understanding a model's capabilities before integrating it. You can see every parameter, whether it's required, what values are valid, and what the defaults are — all without reading external documentation.


Pricing in the Console

Each model's Console page shows transparent pricing information so you know exactly what a run will cost before you execute:

  • Official model badge — indicates the model uses standard Skytells pricing based on computational resources and inference time
  • Cost per unit — shown in the badge row (e.g., $0.03 / image)
  • Cost breakdown — a pricing section shows the output cost and gives a concrete estimate (e.g., "generating 100 images should cost around $3")
  • Learn more link — navigates to the full pricing page for the model

Always check the cost badge before running a model. Pricing varies by model — some charge per image, per token, per second of compute, or per API call.


Walkthrough: Your First Console Run

Here's a step-by-step guide to running your first model in the Console:

Open the Console

Go to skytells.ai/dashboard/console and select a model — for example, TrueFusion for image generation.

Verify Your API Key

Your API key should be pre-filled in the top bar. If it's not, paste your key into the API Key field. Make sure you have at least $2 in credit.

Enter Your Prompt

In the Playground tab, type a description into the Prompt field. For TrueFusion, try something like:

A serene Japanese garden with cherry blossoms and a stone bridge, soft morning light

Adjust Parameters (Optional)

Click Advanced Options to fine-tune the output:

  • Set Aspect Ratio to 16:9 for a widescreen image
  • Increase Number of Images to 4 to compare variations
  • Leave Prompt Optimizer enabled to let the model enhance your prompt

Click Run

Click the ▶ Run button. The model processes your request and results appear in the Output panel within seconds.

View Results

Switch between Preview (visual output), JSON (raw response), and Logs (timing/debug info). If you're happy with the result, switch to the Node.js or Python tab on the input side to grab the code and use it in your application.


Tips for Getting the Most Out of the Console

  • Start with the Form view — it validates inputs, shows required fields, and prevents malformed requests
  • Use the code tabs to bridge from experimentation to production — prototype in the form, copy the generated Node.js/Python/cURL code into your app
  • Check the Examples tab before writing your own prompts — see what inputs produce great results for each model
  • Compare outputs by running the same prompt with different parameters (e.g., different aspect ratios or image counts)
  • Monitor the Logs tab for timing, request IDs, and debug information if something doesn't look right
  • Switch models instantly using the dropdown — no need to navigate away from the Console
  • Use the Schema tab to understand what a model accepts before you build around it

How is this guide?

On this page