Promptchan API Guide Authentication Image Requests Costs Safety and Testing

Direct answer: Promptchan publicly documents an external image-generation API. Create an API key in account settings, authenticate with the x-api-key header and send a POST request to /api/external/create on the documented base service. The guide states that a default image costs one Gem, with some higher-quality options adding cost. Keep the key server-side.

Documented API facts

Item	Official public guide
Current documented use	Image generation
Base URL	`https://prod.aicloudnetservices.com/`
Endpoint	`POST /api/external/create`
Authentication	`x-api-key` request header
Default stated cost	One Gem per image
Added costs	Higher quality and features such as face restoration
Example output	Base64 image plus Gem balance field

Affiliate disclosure: We may earn a commission through the Promptchan partner link. It is not an API discount code. This guide interprets Promptchan's public documentation and does not claim a private production load test.

Open Promptchan and Check API Access

What can the Promptchan API do?

The accessible official guide documents image generation with style, pose, filter, prompt, quality, creativity, size, negative prompt and face-restoration fields.

Do not assume every consumer feature has a matching public endpoint. Promptchan offers chat and video in its main product, but the public getting-started page used for this article does not document chat, live calls, video generation or Extend endpoints.

If a business needs one of those capabilities, contact Promptchan through its Business/API channel and obtain current documentation, limits and written terms before planning the integration.

This page owns developer access. The Promptchan image generator guide covers the no-code interface, while Promptchan Gems owns cross-product budgeting.

Before funding development, compare the current Promptchan pricing plans and review Promptchan safety, privacy and content rights. An API proof of concept should not begin until both billing and data handling fit the application.

How to get a Promptchan API key

Create or sign in to a Promptchan account.
Open account Settings.
Find the API Key section.
Select Generate.
Copy the key once into a secrets manager.
Do not paste it into browser JavaScript, a public repository or a screenshot.
Buy or allocate enough Gems for a small test.
Send one neutral development request.
Confirm the response and Gem balance.
Rotate the key if it is ever exposed.

Use HowTo schema only if the published article keeps this complete visible sequence and the editor rechecks the settings label.

Your first image request

Promptchan's guide provides a request structure similar to this safe example:

curl --request POST \
  'https://prod.aicloudnetservices.com/api/external/create' \
  --header 'Content-Type: application/json' \
  --header 'x-api-key: YOUR_SERVER_SIDE_API_KEY' \
  --data '{
    "style": "Real",
    "poses": "Default",
    "filter": "Cyberpunk",
    "detail": 0,
    "prompt": "Portrait of an original fictional adult traveller in a neon city",
    "seed": 0,
    "quality": "Ultra",
    "creativity": 50,
    "image_size": "512x768",
    "negative_prompt": "rain",
    "restore_faces": false
  }'

The example keeps the subject fictional and adult. It does not copy a real person's identity.

Do not paste a real key into documentation, analytics or support screenshots. Replace it with a placeholder first.

What do the request fields mean?

Field	Purpose	Testing note
`style`	Main visual style	Start with one documented option
`poses`	Pose selection	Keep Default for baseline
`filter`	Additional visual treatment	Test after baseline
`detail`	Detail control	Change one level at a time
`prompt`	Main image instruction	Validate length and prohibited content
`seed`	Repeatability control	Store with each result
`quality`	Output quality	May add Gem cost
`creativity`	Prompt freedom	Benchmark at fixed values
`image_size`	Output dimensions	Allow only supported values
`negative_prompt`	Unwanted details	Keep concise
`restore_faces`	Face correction	Official guide says added features may cost more

Because the public article dates from 2023, field values may evolve. Treat the current API docs and an actual development response as the authority.

How should an application store the response?

Promptchan's example response contains a Base64-encoded image and a gems value. A production service should:

validate the HTTP status and content type;
set a maximum response size;
decode Base64 on the server;
verify the resulting file signature;
generate a random internal filename;
store the prompt and settings separately with restricted access;
record the returned Gem balance; and
remove temporary data after the retention period.

Never insert unvalidated Base64 directly into an unrestricted data pipeline.

API key security

The key authenticates paid requests. Treat it like a password.

Store it in an environment secret or managed vault.
Send requests from your server, not a public browser bundle.
Never log request headers.
Use separate keys for development and production if Promptchan supports it.
Rotate after staff changes or suspected exposure.
Restrict who can view billing and generation logs.
Add a daily Gem ceiling in your own application.
Reject automated retry loops.

Promptchan's terms also prohibit screen scraping for commercial purposes. Use the documented API, not browser automation around the Create page.

Cost and Gem controls

The official API guide says a default image costs one Gem and that higher-quality settings or added features can increase the cost. Do not hard-code 1 Gem as a universal accounting rule.

Use an internal ledger:

Field	Why it matters
Request ID	Trace one generation
User or project ID	Attribute spend
Timestamp	Find incidents
Settings hash	Group equivalent tests
Balance before	Detect unexpected use
Balance returned	Calculate actual deduction
Status	Complete, rejected or failed
Output accepted	Calculate cost per useful result

Set limits at three levels: per request, per user per day and total account per day. Alert before the balance is empty.

Error handling without double spending

Use cautious retries. A network timeout does not prove that generation failed. The server may have accepted the job while the client lost the response.

Before retrying:

record the exact time and payload hash;
inspect any request identifier;
check the Gem balance if available;
use an idempotency feature only if the current docs support it; and
wait for Promptchan's recommended interval.

If the public API lacks request-status or idempotency documentation, ask Promptchan for production guidance. Do not invent an idempotency header.

Handle status classes:

400: validate fields and supported values.
401 or 403: stop and check the key or access rights.
429: back off and respect the returned rate information.
500 range: pause, log and retry only under a capped policy.

These are standard HTTP interpretations. Confirm Promptchan-specific error bodies in the current docs.

A proper API test plan

Functional tests

one valid neutral prompt;
missing required field;
unsupported size;
invalid API key;
insufficient Gems; and
safe rejection of prohibited input.

Quality tests

Use five fixed prompts across supported styles. Store seed, request, time, output and Gem change. Score prompt match, composition, artefacts and consistency.

Reliability tests

Start with a very small request rate. Measure success, response time, timeout and duplicate-charge risk. Obtain written rate limits before any load test.

Security tests

Confirm the key is absent from browser code, logs, errors, analytics and generated download URLs.

Content moderation and consent

Promptchan's terms require users to be at least 18 and prohibit AI-generated child sexual material, non-consensual deepfakes and impersonation of celebrities, public figures or private people without consent.

An API integration needs its own safeguards before sending requests:

adult-only access controls;
clear prohibited-content rules;
input screening;
abuse reporting;
output review before publication;
audit records with restricted access; and
a fast removal process.

Do not rely on the upstream service as the application's only safety layer.

Production readiness checklist

Current business/API permission confirmed.
Endpoint and schema checked against live docs.
API key stored server-side.
Rate limits obtained in writing.
Gem prices and top-up terms verified.
Timeouts and capped retries implemented.
Moderation and age controls active.
User prompts and images protected.
Removal and support contacts documented.
Billing alerts tested.
Terms reviewed for the intended commercial use.

Is the Promptchan API right for your project?

It fits a server-side application that needs Promptchan image generation and can manage Gem-based billing. It is not yet a documented answer for someone who assumes every Promptchan chat and video feature is available programmatically.

Run a small proof of concept. Confirm output, cost, limits, content rights and support before building the whole product around one older getting-started article.

Frequently asked questions

Does Promptchan have an API?

Yes. Promptchan publicly documents an external image-generation API.

Where do I get a Promptchan API key?

The official guide says to generate one from the API Key section in Settings.

How do I authenticate?

Send the key in the x-api-key request header.

What is the Promptchan API endpoint?

The public guide lists POST /api/external/create on https://prod.aicloudnetservices.com/.

How much does one API image cost?

The guide states one Gem by default, with added cost for some settings.

Does the API support Promptchan video or chat?

Those endpoints were not documented in the public getting-started page used here. Ask Promptchan for current business documentation.

Can I expose the key in a mobile or web app?

No. Route paid requests through a controlled server and keep the key secret.