- Monthly visits
- 87,998
- 2026-09
- Monthly growth +1.3%
Modelflare - Unified OpenAI-compatible gateway for 70+ AI models across 13 providers
Modelflare is an OpenAI-compatible AI API gateway that gives developers and teams unified access to over 70 text, image, and video models from 13 different providers. It supports smart provider routing, automatic upstream fallback, request-level usage and cost tracking, and a no-code creative workbench for image and video generation. The platform serves engineering teams building AI applications, creative studios, individual developers, and enterprise teams with custom workload requirements.
- AI DevTools
- Model APIs
- Software
- GPT
- Claude
- Gemini
- API Available
- Enterprise
2Views
Capabilities
OpenAI-compatible unified API gateway
Provides a single OpenAI-compatible endpoint to access over 70 text, image, and video models from 13 different providers, eliminating separate API integrations for each service.
Smart routing and automatic fallback
Routes requests using balanced, stability-first, or low-price-first strategies, and automatically retries eligible requests on alternate routes for the same model if an upstream provider fails.
Request-level usage and cost tracking
Logs individual request metadata including model used, token counts, latency, and calculated cost, and displays all records in the user dashboard for full visibility.
Per-project API key management
Creates separate API keys for each project, with configurable usage quotas, expiration dates, and model or IP access restrictions to isolate workloads.
Castaly no-code creative workbench
Offers a browser-based interface for selecting image and video generation models, entering prompts, and generating media without writing API code, using the same shared account balance.
Asynchronous video model task management
Creates video generation tasks via API, returns a task ID for polling, and lets users download completed video results once processing finishes.
Pricing
Modelflare offers paid pay-as-you-go top-up billing and tiered 28-day subscription plans. Compare included credit bonuses, supported model coverage, and total credit value when selecting a plan.
| Plan | Price | Included features and limits |
|---|---|---|
| $99.99 28-day subscription | $99.99 28-day subscription |
|
| $299.99 28-day subscription | $299.99 28-day subscription |
|
| $499.99 28-day subscription | $499.99 28-day subscription |
|
| $999.99 28-day subscription | $999.99 28-day subscription |
|
| $1499.99 28-day subscription | $1499.99 28-day subscription |
|
A single top-up of US$20 or more unlocks the discounted openai-first-topup routing group, while a single top-up of US$50 or more unlocks the discounted claude-first-topup routing group. Text models are billed per million input and output tokens, image models per image or token, and video models per second or token.
Use cases
Build multi-model AI applications
You are developing an AI-powered application that needs to call multiple LLM, image, and video models. You integrate the single Modelflare API endpoint, configure routing rules, and avoid maintaining separate integrations for each provider.
Track and control per-project AI costs
You run multiple independent AI projects with separate teams. You create dedicated API keys for each project with custom quotas, and view isolated usage and cost records to prevent unexpected overspend.
Generate creative image and video content
You are a creative professional testing different generative media outputs. You use the Castaly no-code workbench to experiment with dozens of image and video models, compare results, and generate final assets without writing code.
Deploy production AI workloads with reliability
You run production AI services that require consistent uptime. You configure ordered backup routes for critical models, enable automatic upstream fallback, and reduce service disruption from individual provider outages.
Limitations
All video model operations are asynchronous and require polling to retrieve completed results.
The Request Archive diagnostic feature is disabled by default for all accounts.
Modelflare does not store full payment card numbers, with all payment processing handled by third-party providers.
Different models support different API endpoints, parameters, and protocols that are not universally compatible.
Unused subscription credits expire at the end of each 28-day billing period.
The standard Modelflare service does not currently claim SOC 2 or ISO 27001 certification and is not HIPAA-enabled.
Who it is for
Product and engineering teams building AI-powered applications
Creative studios and agencies working with image and video generation models
Individual developers using AI models for development and testing
Enterprise teams with custom workload and capacity requirements
FAQ
Do Castaly generations and API requests share the same balance?
Yes, Castaly uses your Modelflare account and shared wallet for all image and video generation, with unified usage records across both interfaces.
Can every model use the same API format?
No, supported endpoints and parameters depend on the specific model and selected route. Check the official model documentation for exact required inputs and supported protocols.
How are image and video requests billed?
Billing rules vary by model, with charges calculated per image, per second of video, or per token. Resolution, duration, and reference inputs may affect the final cost.
What happens when an upstream provider fails?
When a failure allows failover, Modelflare automatically routes the request to another eligible backup route for the same model, using the same API key you configured.
Do I need a subscription to start using Modelflare?
No, you can start with pay-as-you-go credit top-ups. Subscriptions are optional for users who prefer predictable 28-day billing with bonus credits.
Comments