AI Use Policy

01

What this policy covers

Muqira AI Studio, Muqira Flow, Scriptology, and any other feature that generates, summarises, transforms, or acts on content using a third-party AI model (collectively, "AI Features") are governed by this AI Use Policy in addition to our Terms of Service.

02

How Muqira routes your requests

Muqira does not train or operate our own foundation models. AI Features work by routing your request to third-party AI providers, which currently may include Anthropic, OpenAI, Google (Gemini), Groq, Mistral, DeepSeek, and models available via OpenRouter, depending on your plan, the feature you're using, and provider availability at the time. Which specific provider and model handles a given request can change — due to your plan tier, an outage or rate limit on one provider, or ongoing product changes. Where the interface displays which model served a response, that reflects the actual provider used for that request. If you supply your own API key for a provider ("Bring Your Own Key" / BYOK), requests using that key are sent using your credentials and are subject to that provider's own terms and usage policies, not ours.

03

Data sent to AI providers

To generate a response, the content of your request (and, where relevant, attached files, conversation history, or workflow context) is sent to the selected AI provider's API. This is necessary for the feature to function — there is no way to generate an AI response without sending the input to a model. We do not sell this data. Whether a given provider uses API input to train its own models is governed by that provider's terms, not ours — see "Training use" below. We select providers whose standard API terms state that API-submitted data is not used for training by default, but you should not rely solely on our summary; each provider's own current terms are the authoritative source, and provider policies can change without notice to us.

04

Training use — what we can and cannot promise

Muqira does not use your content to train our own models, because we do not operate foundation models. We do not promise, on behalf of any third-party AI provider, that your data is never used for training. Some providers offer contractual or account-level guarantees against training use (particularly for paid API tiers); others may differ. Where you have specific requirements here (e.g. for regulated data), contact legal@muqira.com before relying on any AI Feature for that use case, and consider using your own API key with a provider whose terms you have independently verified.

05

Accuracy and hallucination risk

AI-generated content can be factually incorrect, outdated, biased, or entirely fabricated ("hallucinated"), including confident-sounding citations, statistics, or quotations that do not exist. This is an inherent limitation of current AI models, not a defect specific to Muqira. You are responsible for reviewing, fact-checking, and editing AI-generated content before relying on it or publishing it — including content generated through automated Muqira Flow workflows that publish without a human-approval step.

06

Do not use AI Features for high-impact decisions

Muqira's AI Features are not designed, tested, or intended for use in making or materially influencing decisions about an individual's: • Employment, hiring, or termination • Credit, lending, or insurance eligibility • Medical diagnosis or treatment • Legal rights, eligibility for government benefits, or immigration status • Law enforcement or criminal justice outcomes Do not use Muqira's AI Features as the basis for these categories of decisions. Where any automated processing through Muqira produces an effect on an individual that is legally or similarly significant, applicable law (including GDPR Article 22 and India's DPDP Act) may require meaningful human review before that outcome is acted on — that review is your responsibility as the operator of your workflows.

07

Copyright and generated content

You are responsible for ensuring you have the right to use any input you provide to an AI Feature (e.g. text, images, or files you upload) and for how you use the output. AI-generated output may not be eligible for copyright protection in every jurisdiction, and the same or similar output could in principle be generated for another user from a similar prompt. Muqira makes no representation that generated content is unique, non-infringing, or eligible for copyright registration. Do not use AI Features to reproduce copyrighted material you don't have rights to, or to generate content that infringes a third party's trademark or publicity rights.

08

Provider outages and fallback

If your selected or default AI provider is unavailable, Muqira's AI Router may automatically retry with a fallback provider/model appropriate to your plan tier, so your request can still complete. This means the exact model that serves a given request is not always guaranteed in advance. If you require a specific model with no fallback, contact support@muqira.com to discuss available options for your plan.

09

Changes to this policy

We may update this policy as we add AI providers, features, or as provider terms change. Material changes will be communicated the same way as changes to our Terms of Service.

10

Contact

Email: legal@muqira.com General: hello@muqira.com Muqira Technologies Mumbai, Maharashtra, India