Skip to content
Model infoModelsSnapshot: 15 Oct 2026

Magnific / formerly Freepik / Mystic

Mystic API

Hosted image-generation workflow combining underlying model and enhancement stages, with six exposed configurations and conditional LoRA support.

Image generation
availableClosed-source managed componentImage generation
Official documentation (opens in a new tab)Reviewed 15 Oct 2026
6 of 20 specification fields collected
Output resolution1k, 2k, 4kSee specification notes below
Output formatsNot collectedAwaiting source review
Input billingNot collectedMagnific API
Output billingNot collectedNo inferred price

Specifications

Provider-documented facts. “Not collected” means we have not verified that field yet, not that the feature is unsupported.

Context / input limit
Not collected

Not established for this exact model or operation. Media controls, research training sizes and integration protocols do not establish language-model token or capability limits.

Maximum input tokens
Not collected

Not established for this exact model or operation. Media controls, research training sizes and integration protocols do not establish language-model token or capability limits.

Maximum output
Not collected

Not established for this exact model or operation. Media controls, research training sizes and integration protocols do not establish language-model token or capability limits.

Maximum prompt characters
Not collected

Not established for this exact model or operation. Media controls, research training sizes and integration protocols do not establish language-model token or capability limits.

Reasoning
Not collected

Not established for this exact model or operation. Media controls, research training sizes and integration protocols do not establish language-model token or capability limits.

Tool calling
Not collected

Not established for this exact model or operation. Media controls, research training sizes and integration protocols do not establish language-model token or capability limits.

Structured output
Not collected

Not established for this exact model or operation. Media controls, research training sizes and integration protocols do not establish language-model token or capability limits.

Streaming
Not collected

Not established for this exact model or operation. Media controls, research training sizes and integration protocols do not establish language-model token or capability limits.

Batch processing
Not collected

Not established for this exact model or operation. Media controls, research training sizes and integration protocols do not establish language-model token or capability limits.

Fine-tuning
SupportedSource for Fine-tuning (opens in a new tab)

Hosted style and character LoRA training. Compatibility is restricted by the generation configuration; weights export is not established.

Context caching
Not collected

Not established for this exact model or operation. Media controls, research training sizes and integration protocols do not establish language-model token or capability limits.

Knowledge cutoff
Not collected

Not established for this exact model or operation. Media controls, research training sizes and integration protocols do not establish language-model token or capability limits.

Release date
Not collected

Not established for this exact model or operation. Media controls, research training sizes and integration protocols do not establish language-model token or capability limits.

Parameters
Not collected

Not established for this exact model or operation. Media controls, research training sizes and integration protocols do not establish language-model token or capability limits.

Architecture
Provider describes a workflow incorporating Flux, Stable Diffusion and Magnific technology. Exact private derivatives and version mappings are not published here.Source for Architecture (opens in a new tab)
Languages
Not collected

Not established for this exact model or operation. Media controls, research training sizes and integration protocols do not establish language-model token or capability limits.

Input modalities
Text, ImageSource for Input modalities (opens in a new tab)
Output modalities
ImageSource for Output modalities (opens in a new tab)
Output resolutions
1k, 2k, 4kSource for Output resolutions (opens in a new tab)

Default 2k. Resolution labels do not establish native base-model dimensions.

Generation modes
realism (default), fluid, zen, flexible, super_real, editorial_portraitsSource for Generation modes (opens in a new tab)

Configuration labels are not six proven independent foundation models.

Access & pricing

Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.

Magnific API

Provider source (opens in a new tab)
API model ID
realism | fluid | zen | flexible | super_real | editorial_portraits

The model field is a workflow configuration, not a checkpoint ID.

Access
Documented credit-metered API. API key and account eligibility required; no authenticated generation was attempted.
Regions
Deployment regions not established
API endpoint URL
POST https://api.magnific.com/v1/ai/mysticSource for API endpoint URL (opens in a new tab)
Authentication
x-magnific-api-key header; JSON request body.Source for Authentication (opens in a new tab)
Request quotas
Published limits: 50 requests/second over 5 seconds per IP; 10/second average over 2 minutes per IP; 30,000/minute per API key. Account agreements and response headers may impose additional limits. No daily generation cap is listed for these routes.Source for Request quotas (opens in a new tab)
Billing scope
AI API calls consume credits on every paid plan. Web unlimited allowances do not cover API calls. Exact per-operation API tariffs and web-version mappings remain unverified.Source for Billing scope (opens in a new tab)
Required request fields
No required fields listed in this request schema.Source for Required request fields (opens in a new tab)
API parameters
View all 15 entries
  • prompt: string
  • webhook_url: string
  • structure_reference: string
  • structure_strength: integer ≥ 0 ≤ 100; default 50
  • style_reference: string
  • adherence: integer ≥ 0 ≤ 100; default 50
  • hdr: integer ≥ 0 ≤ 100; default 50
  • resolution: one of 1k, 2k, 4k; default 2k
  • aspect_ratio: one of square_1_1, classic_4_3, traditional_3_4, widescreen_16_9, social_story_9_16, smartphone_horizontal_20_9, smartphone_vertical_9_20, standard_3_2, portrait_2_3, horizontal_2_1, vertical_1_2, social_5_4, social_post_4_5'; default square_1_1
  • model: one of realism, fluid, zen, flexible, super_real, editorial_portraits; default realism
  • creative_detailing: integer ≥ 0 ≤ 100; default 33
  • engine: one of automatic, magnific_illusio, magnific_sharpy, magnific_sparkle; default automatic
  • fixed_generation: boolean; default false
  • filter_nsfw: boolean; default true
  • styling: object
Source for API parameters (opens in a new tab)

Literal schema fields; numeric ranges are endpoint constraints, not model training specifications.

Asynchronous generation
Returns a task ID for polling or optional webhook completion. Task updates do not establish token streaming.Source for Asynchronous generation (opens in a new tab)
Reference images
One Base64 structure reference and one Base64 style reference; either causes LoRAs to be ignored. structure_strength, adherence and hdr: 0–100, default 50.Source for Reference images (opens in a new tab)
LoRA compatibility
Documented compatible path: omit model and both reference images. fluid, flexible, super_real and editorial_portraits ignore LoRAs; references also disable them. Incompatible LoRAs are silently ignored rather than rejected. Zen compatibility is not explicitly established.Source for LoRA compatibility (opens in a new tab)
Style and character controls
At most one style and one character entry; each strength 0–200, default 100. Color palette has 1–5 entries.Source for Style and character controls (opens in a new tab)
Aspect ratios
View all 13 entries
  • square_1_1
  • classic_4_3
  • traditional_3_4
  • widescreen_16_9
  • social_story_9_16
  • smartphone_horizontal_20_9
  • smartphone_vertical_9_20
  • standard_3_2
  • portrait_2_3
  • horizontal_2_1
  • vertical_1_2
  • social_5_4
  • social_post_4_5'
Source for Aspect ratios (opens in a new tab)

Fluid only supports 1:1, 9:16, 16:9, 3:4 and 4:3. The schema's social_post_4_5' has a trailing apostrophe; it is retained as an unresolved contract defect.

Seed controls
fixed_generation defaults false; provider says identical settings produce a consistent image when true. No explicit seed field or billed cache contract.Source for Seed controls (opens in a new tab)

Endpoint pricing has not been collected. Check the provider for applicable rates.

Data & enterprise terms

General terms link separate AI and API agreements. Exact endpoint retention, training use and residency are not established by this review.

Read the service policy (opens in a new tab)

Magnific API

Provider source (opens in a new tab)
API model ID
Not collected

Exact checkpoint selector is not established.

Access
Documented credit-metered API. API key and account eligibility required; no authenticated generation was attempted.
Regions
Deployment regions not established
API endpoint URL
POST https://api.magnific.com/v1/ai/loras/stylesSource for API endpoint URL (opens in a new tab)
Authentication
x-magnific-api-key header; JSON request body.Source for Authentication (opens in a new tab)
Request quotas
Published limits: 50 requests/second over 5 seconds per IP; 10/second average over 2 minutes per IP; 30,000/minute per API key. Account agreements and response headers may impose additional limits. No daily generation cap is listed for these routes.Source for Request quotas (opens in a new tab)
Billing scope
AI API calls consume credits on every paid plan. Web unlimited allowances do not cover API calls. Exact per-operation API tariffs and web-version mappings remain unverified.Source for Billing scope (opens in a new tab)
Required request fields
images, name, qualitySource for Required request fields (opens in a new tab)
API parameters
name: string, description: string, quality: one of medium, high, ultra, images: array; minimum 6 items; maximum 20 items, webhook_url: stringSource for API parameters (opens in a new tab)

Literal schema fields; numeric ranges are endpoint constraints, not model training specifications.

Asynchronous generation
Returns a task ID for polling or optional webhook completion. Task updates do not establish token streaming.Source for Asynchronous generation (opens in a new tab)
Training inputs
Style LoRA: 6–20 image URLs; medium, high or ultra quality.Source for Training inputs (opens in a new tab)

Endpoint pricing has not been collected. Check the provider for applicable rates.

Data & enterprise terms

General terms link separate AI and API agreements. Exact endpoint retention, training use and residency are not established by this review.

Read the service policy (opens in a new tab)

Magnific API

Provider source (opens in a new tab)
API model ID
Not collected

Exact checkpoint selector is not established.

Access
Documented credit-metered API. API key and account eligibility required; no authenticated generation was attempted.
Regions
Deployment regions not established
API endpoint URL
POST https://api.magnific.com/v1/ai/loras/charactersSource for API endpoint URL (opens in a new tab)
Authentication
x-magnific-api-key header; JSON request body.Source for Authentication (opens in a new tab)
Request quotas
Published limits: 50 requests/second over 5 seconds per IP; 10/second average over 2 minutes per IP; 30,000/minute per API key. Account agreements and response headers may impose additional limits. No daily generation cap is listed for these routes.Source for Request quotas (opens in a new tab)
Billing scope
AI API calls consume credits on every paid plan. Web unlimited allowances do not cover API calls. Exact per-operation API tariffs and web-version mappings remain unverified.Source for Billing scope (opens in a new tab)
Required request fields
gender, images, name, qualitySource for Required request fields (opens in a new tab)
API parameters
name: string, description: string, quality: one of high, ultra, gender: one of male, female, neutral, custom, images: array; minimum 8 items; maximum 20 items, webhook_url: stringSource for API parameters (opens in a new tab)

Literal schema fields; numeric ranges are endpoint constraints, not model training specifications.

Asynchronous generation
Returns a task ID for polling or optional webhook completion. Task updates do not establish token streaming.Source for Asynchronous generation (opens in a new tab)
Training inputs
Character LoRA: 8–20 image URLs; high or ultra quality. Gender field required: male, female, neutral or custom.Source for Training inputs (opens in a new tab)

Endpoint pricing has not been collected. Check the provider for applicable rates.

Data & enterprise terms

General terms link separate AI and API agreements. Exact endpoint retention, training use and residency are not established by this review.

Read the service policy (opens in a new tab)

Caching

Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.

01 / INITIAL INPUTProcess & retain

The system processes eligible input and retains reusable state.

02 / LATER REQUESTCheck available state

Reuse depends on matching input, session state and the cache policy.

03 / CACHE HITReuse & generate

Reusing state can reduce repeated work. Any charges follow the billing rules below.

Magnific API

realism | fluid | zen | flexible | super_real | editorial_portraits
Not verified

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

How it works
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Minimum reusable input
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Lifetime & refresh
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

How to enable it
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Usage fields
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Billing rules
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Limits & data boundaries
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 15 Oct 2026

Magnific API

Not verified

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

How it works
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Minimum reusable input
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Lifetime & refresh
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

How to enable it
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Usage fields
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Billing rules
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Limits & data boundaries
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 15 Oct 2026

Magnific API

Not verified

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

How it works
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Minimum reusable input
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Lifetime & refresh
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

How to enable it
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Usage fields
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Billing rules
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

Limits & data boundaries
Not collected

No reviewed prompt-cache support contract or tariff. Fixed generation, image history and asynchronous task polling are not prompt caching.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 15 Oct 2026

What the measurements say

No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.

Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.

The record behind the model

Why it is included

Managed image/video workflow with a documented hosted interface. The exact constituent checkpoints and any private derivative weights are not established. This record is not counted as a distinct closed-weight foundation model.

Classification evidence (opens in a new tab)

Sources reviewed