Common Questions

On this page, we will answer some common questions we often get asked, along with other general questions we think you should know the answers to.

Common questions about products and services

APIs:

When will API limits reset?

API limits reset on the first of each month. This allows for constant uninterrupted access to our APIs. For example: if you have started your API usage on the 15th of August, your call limits will still reset on the first of each month, meaning, the first of September.

Important note for premium accounts: If you downgrade from a premium plan, your existing API calls from the premium account will remain available until the next billing cycle (1st of the month) before being reset to your current plan's limits.

When you reach your API limit, or you don't have any more calls available, what will happen?

When you reach your API call limit, limited either by rate or you don't have any more calls available, the API will return an error, stating either of the 2 cases. All API errors are billable, including rate limits and other error codes that reach the backend, since the API resources have been accessed.

Calling the API multiple times after being rate-limited, may result in temporary or permanent action against your account.

Can APIs be called from multiple programming languages?

Yes, our APIs work cross-platform and across a variety of programming languages. Just be sure to pass the correct parameters to the API in order to get a successful API call.

If I initiate a normal API call and it fails due to an error in my implementation, does it cost an API call?

Yes, if the API is live and working, with no errors on our end, it will cost an API call since you have invoked the API. Errors such as implementation errors, like failure to call the actual API, will not cost an API call. Parameter errors like missing or invalid parameters does cost an API call, since it still calls the API.

This is why we created Developer API calls, since they can help you to quickly debug your implementation, before running in production environments.

Are there free API calls?

Yes, all accounts are registered with the free plan upon sign-up. With the free plan, you get free monthly API calls, that you can use on various APIs, across our platform. You can increase your limits either by purchasing a premium plan, or by purchasing more calls. API calls reset on the first of each month.

How do I authenticate my API requests?

Authentication can be done in a few ways. The preferred method is to use a Bearer token in the Authorization header with your Developer or standard API token. Alternatively, you can pass the token in the `api_key` parameter for POST requests, or the `token` parameter for GET requests. For easier integration and testing, you can also use the OpenAPI specification files available for each API.

What are the common error codes and what do they mean?

Here are some of the common error codes:

• 400: Missing or invalid parameters. This occurs when the parameters you passed do not match what the API expects.

• 401: Error during the request. This can be due to insufficient API calls, input errors, or errors from the API's internal logic.

• 405: Invalid method. This means you are using an incorrect HTTP method (e.g., GET instead of POST) to access the API.

• 429: Rate limit exceeded. You have made too many requests in the given timeframe.

Remember: All of these errors are billable since they reach our backend systems.

Is there a changelog for the APIs to keep track of updates and new features?

Yes, a changelog is available at the top of each API's documentation. For an example, see [here](https://infinitode-docs.gitbook.io/documentation/api-documentation/basic-math-api-documentation-experimental).

How do I get started with your APIs?

Getting started guides and detailed API documentation are available on our [GitBook documentation page](https://app.gitbook.com/o/5JUG1pl8gcE8caqjeo9B/s/lK1OH5154DcjEyLuJPtz/). The typical process is: Sign up → Go to account settings (auto-redirected after signup) → Copy your API key (Developer or normal) → Start making API calls using our documentation.

What's the difference between different API types (experimental, standard, etc.)?

We offer different types of APIs based on their development stage and reliability:

• Standard APIs: Production-ready APIs that have met our quality, accuracy, and reliability standards.

• Experimental APIs: APIs that run on our infrastructure but haven't met the mark in terms of goal, accuracy, or reliability. These may produce outputs that are irrelevant, misleading, or incorrect at times and are not intended for production environments.

• Developer APIs: Testing versions of our APIs that return dummy data for integration testing without using paid API calls.

Do you support both GET and POST methods for API calls?

Yes, our APIs support both GET and POST methods. We also provide OpenAPI specification files for easier integration and testing, available at `https://infinitode.netlify.app/openapi/api-name/openapi.yaml` (replace "api-name" with the specific API you're working with).

How is my data handled and secured?

All data is securely handled by Netlify and transmitted directly via HTTPS to our serverless APIs. Your API keys and all user data are encrypted at rest to ensure maximum security and privacy.

Can I integrate your APIs into my existing business workflow?

Absolutely! Our APIs are designed to be business-friendly and can be integrated into various workflows. Whether you're building customer-facing applications, internal tools, or automated processes, our APIs provide the flexibility and reliability businesses need. Check our documentation for specific use cases and integration examples.

What happens if an API I'm using gets deprecated?

We provide advance notice before deprecating any APIs. When an API is scheduled for deprecation, we'll notify all users via email and update the API documentation with migration guides and timelines. We strive to provide alternative solutions and sufficient transition time for businesses relying on deprecated APIs.

OPEN-ARC:

What is OPEN-ARC?

OPEN-ARC is a collaborative AI research project where we release basic implementations, and sometimes base models to solve certain 'challenges'. It is beginner-friendly, and we host leaderboards for each project, driving users to outperform the previous implementation.

Are the base models free to use?

Personally yes, commercially no. Other users' models, implementations, and code inside of our OPEN-ARC project are subject to the copyright license of their respective owners/maintainers/developers. Our models, implementations and code however, are subjected to the license found at: https://github.com/Infinitode/OPEN-ARC/blob/main/LICENSE.md

How often are projects added to OPEN-ARC?

We try our best to release a new project each second week, but it may sometimes take a bit longer.

How can I contribute to an OPEN-ARC project?

To contribute, you can submit a pull request on our GitHub repository. In your pull request, please include your changes to the leaderboard file and a publicly accessible link to your notebook or project. You can include model files, but it's not a requirement.

What are the judging criteria for the leaderboards?

The main judging criterion is the accuracy score on the test set. In some cases, such as for text-based challenges, the BLEU score is used. It's important to note that these scores are primarily an indication of model performance on a specific dataset and not judged too heavily.

Can I use OPEN-ARC models in my commercial projects?

It depends on the specific model and its license. Our models are subject to the license found in our repository, but other contributors' models may have different licensing terms. Always check the specific license for each model before using it commercially. When in doubt, contact us for clarification.

How do I stay updated on new OPEN-ARC challenges and releases?

The best way to stay updated is to watch our GitHub repository for new releases and follow our announcements. You can also check the repository regularly as we aim to release new challenges every two weeks.

What types of AI challenges are featured in OPEN-ARC?

OPEN-ARC features a variety of AI challenges spanning different domains including natural language processing, computer vision, machine learning, and experimental AI research. Each challenge is designed to be beginner-friendly while still providing opportunities for advanced researchers to contribute innovative solutions.

Account & Billing:

How do I upgrade or downgrade my plan?

You can manage your subscription through your account settings. When upgrading, the new limits take effect immediately. When downgrading, your current API calls remain available until the next billing cycle (1st of the month), after which your limits will reset to the new plan's allocation.

Can I purchase additional API calls without upgrading my plan?

Yes, you can purchase additional API calls as add-ons to your current plan. These additional calls follow the same monthly reset cycle as your base plan allocation.

How does billing work for API usage?

Billing is based on successful API calls within your monthly allocation. Calls that fail due to network issues before reaching our backend are not billed, but any call that reaches our servers (including those that return error codes) counts toward your usage.

Can I get a refund for unused API calls?

API calls reset monthly and do not roll over to the next billing period. We do not provide refunds for unused calls, but you can adjust your plan for the next billing cycle to better match your usage needs.

Troubleshooting:

My API calls are failing with authentication errors. What should I check?

First, verify that you're using the correct API token and that it hasn't expired. Check that you're passing the token correctly (Bearer token in Authorization header, or in the appropriate parameter for GET/POST requests). If you suspect your token is compromised, contact support immediately for a reset.

I'm getting timeout errors when calling the API. What could be the cause?

Timeout errors can occur due to network issues, high server load, or complex processing requirements. Try reducing the complexity of your request, implement retry logic with exponential backoff, or contact support if the issue persists across multiple API endpoints.

How can I debug API integration issues?

Use Developer API tokens to test your integration without consuming paid API calls. These tokens return dummy data in the same format as live APIs, allowing you to verify your implementation before going live. Also, check the OpenAPI specification files available for each API for proper parameter formatting.

Why am I getting unexpected results from AI APIs?

AI APIs can sometimes produce unexpected outputs due to the nature of machine learning models. Check that your input parameters are properly formatted and within expected ranges. If you're using experimental APIs, be aware that they may produce less reliable results. For consistent issues, provide specific examples when contacting support.

---

Looking for the Docs? They've moved to here...