Privacy policy

Last updated 2026-10-15

Thommidhi ("we", "us") is a behaviour simulation platform. It helps people test decisions against simulated populations and then checks those predictions against what actually happened.

This policy explains what information we collect, why, where it goes, how long we keep it and what you can do about it. It covers thommidhi.com and the Thommidhi application.

Who we are

Thommidhi is an early-stage company based in India. We launched recently and are operating in stealth, so we don't publish details of our team or legal entity on this website.

If you need information about the company or entity behind Thommidhi, for example for a contract, a data-protection request or a verification, email hello@thommidhi.com and we'll provide it.

Information we collect

We collect only what the service uses.

Account information: your name, email address, organisation name and password. Passwords are stored as salted one-way hashes, never in readable form. If you sign in with Google, Google gives us your name, verified email address and an account identifier. We never receive your Google password, and signing in with Google gives Thommidhi no access to your YouTube channel.

Session information: a sign-in cookie that keeps you signed in for up to 90 days from when you sign in, after which you sign in again, and a record of that session on our servers. Signing out ends it at once. Brands signing in to the marketplace get a separate cookie that lasts up to 30 days.

If you connect an AI assistant with Oli by Thommidhi: which assistant it was, what you allowed it to do, when you connected it and when it was last used, and a log of which Oli tool ran and when (never what you asked or what came back). Also your choice about sharing your minimum fee, and the plans you save from an assistant or paste back. The next sections describe what an assistant receives.

Simulation data: what you enter, such as decisions, options, audiences, questions, signals and real-world outcomes, plus the results we generate. This includes predictions, confidence assessments and explanations.

What you tell us about your channel: the kind of channel it is, the languages and formats you publish in, the topics you cover, who you believe it is for, your brief, the ideas and treatments you compare, and what you do with the opportunities we show you (saved, dismissed and why).

What you type for your media kit, only if you make one: a line about you, a contact email and WhatsApp number, your Instagram handle and numbers, and your rates. It is used only to build your kit, shown only to you, and never published by us: you download the kit and decide who gets it.

How you work with brands, only if you use Collabs (Beta): what you write about your style and what fits, the categories you welcome and the products you won't promote, the languages and deliverables you offer, your minimum fee, the usage rights, exclusivity and rounds of changes you accept, when you're available, links to example videos you choose, which of these a brand may see, and whether you're open to briefs from brands. Also what you add to your deal desk (your own ideas that could carry a sponsor, brands you're researching and offers you received: brand names, products, fees, dates, terms and notes), the concepts and agreed briefs you write, outcomes you record, and each change of stage with the reason given. All of it is what you typed; none of it is read from YouTube. Offer checks are made by fixed rules from what you entered.

How you appear beyond your Lab, only if you switch it on in Collabs: the name you want shown, a link to your channel, your kind of channel and your size as you tell us (never read from YouTube for this), whether other creators may find you, whether invited brands may see you, and whether your own page for brands is on. Collab suggestions you send or answer: the idea, your note and the contact you choose to share. Ideas you choose to show brands, in your words.

To choose example text and the creators shown to you first, Collabs may use the Indian state your connection comes from, as our hosting provider reports it with each request. It is used for that page only and never stored.

Brands on the private creator marketplace: the brand name, contact name, email and website a brand gives when it makes its page, or that we add when we invite it, one-time sign-in links (we keep only a hash), when they last signed in, and which creators they said they're interested in, with their note. What a brand says it's looking for (its open call): the product, what it'd like made, languages, kinds of content, timing and, if it gives one, a budget range. What we read from those words (topics, languages, who it wants on camera, who it's talking to, each with the words it came from, how and when it was read), and the brand's own correction of it. If someone enters an email at the marketplace that isn't invited, we keep that address and when it was entered, so we can follow up.

Local businesses that join our local spotlight: the business's name, kind, city and area, what it sells and why a creator's viewers would care (in its own words), its links, a contact name and email, a phone number if given, and whether we may name it as a business Thommidhi featured. Joining also makes its page on the private creator marketplace, which waits for our approval before it sees any creator.

Local businesses we list from public sources, which haven't joined Thommidhi: described in full under “Local businesses we list from public sources” below.

If you raise your hand for a brand's open call in Collabs: which call, and the note you add.

If you put your work forward for Thommidhi Spotlight or Creator Honours (your choice; joining Creator Lab never does it): the category, language and link you give, the version of the rules you agreed to, and when you took it back. We ask before showing anything of yours.

Paid briefs: what a brand posts (whether it commissions content or hires a creator to publish it, what to make, where, languages, pay per creator, funded places, for published briefs the smallest audience it wants and how long the post stays live, the rights and terms, a deadline and notes) and our decision on it; for a creator who applies, their short approach, the link to existing work they choose, whether we select them (and any note we give), the link they deliver, the UPI id they give to be paid, our check and the payment we record. The brand sees how many creators applied and, for selected creators only, their name, the work link they chose and the delivered link, never their UPI id or contact details.

Brand briefs, if you send one from our brands page: your name, work email and, if you give it, phone number; your brand's name; and the brief (the product, the customer's problem it solves, what the collaboration should achieve, budget, languages and regions, where it's sold, timeline, deliverables, requirements and no-gos). We keep a one-way hash of your IP address to limit spam, never the address itself, and only a hash of your private status link.

The days you use the Lab: one mark per day, not what you did or looked at, so we can see whether the Lab is useful to you and who might need a hand. It is deleted with your account.

Voice input, only if you choose to speak instead of typing. When you press Speak, your browser asks for the microphone and records up to 30 seconds. The recording is sent to us, turned into text, and discarded. We do not keep the recording or the text made from it, and we never listen in the background: the microphone is on only between pressing Speak and pressing Done.

A video file, only if you choose to add one to see what was happening at the moments viewers left one of your videos. The file is opened in your browser and never uploaded. Your browser measures the sound and the picture itself (loudness, silences, speaking pace, scene cuts) and sends us only short clips of the sound, about 25 seconds around each of those moments and the first 25 seconds of the video, with those measurements. We keep what was found at each moment (for example, a stretch with no one speaking or a sponsor read, with its times), not the file, the clips or the words said.

Early-access requests: the name, email address, YouTube channel or company, and message you send through the early-access form.

Audience poll answers, if a creator shares a poll with you and you answer it: which option you picked, the order the options were shown in and how long you took, with a one-way hash of a random ID kept in your browser (so each person counts once) and a one-way hash of your IP address (to limit floods). We never ask for your name, email address or account, and never store your IP address itself. The creator sees only totals, never who picked what.

Connected platform data, only if you connect a YouTube channel. The next section describes it in full.

Instagram, only if you connect it in Buzz: your Instagram username and account type; the IDs of your recent posts; the text and time of comments on your own posts, of comments and captions where you're @mentioned, and of posts you're tagged in, with links to them. We don't receive or keep the profiles of the people who wrote them. Read with your permission through Instagram's official API. The section “Connected Instagram data” describes it in full.

Job applications, if you apply for a role: the details and answers you submit. We also keep a one-way hash of your IP address and your browser's user-agent to limit spam. We never store your IP address itself.

Messages you send us, for example to our email address. If you send us a phone or WhatsApp number in a message, we keep it with your name and business to reply to you, for example on WhatsApp; it's seen only by our team and you can ask us to delete it at any time.

We don't use advertising or third-party analytics trackers. Our hosting provider, Cloudflare, may set cookies needed to protect the site from abuse. Your light or dark display preference is saved in your own browser, and so is a random ID if you answer a creator's poll, so that you count once.

Connected YouTube data

Connecting a YouTube channel is optional. You authorise it through Google's own sign-in screen, which shows exactly what is being requested. We ask for two read-only permissions:

View your YouTube account. We use it to list your uploads and to record each video's title, thumbnail, publish time, duration, privacy status, category and audio language (as you set them in YouTube Studio) and public view, like and comment counts as they change. The category, audio language, durations and recurring words in your titles are used, by fixed rules and never by an AI model, to suggest what kind of channel yours is, for you to confirm or correct.

View YouTube Analytics reports for your content. We use it to score your Creator Lab tests (a published video's first-week views), to show you growth insights for your own channel (where views come from, subscribers vs people who don't subscribe, audience age and gender, top cities and countries, which videos reached new people, how the last four weeks compare with the four before and which sources changed, the YouTube searches that brought viewers, where viewers stop watching your latest longer videos, the moments they leave fastest or go back to rewatch, and how that compares with videos of similar length on YouTube, subscribers gained and lost per video, how Shorts, longer videos and live compare, and how often YouTube showed your videos and how many of those showings became views, from the daily reach reports), all read by fixed rules, and to start YouTube's daily reach reports, such as impressions and click-through, which only exist from the day they are requested. Growth insights are shown only to your workspace and are never combined with other channels' data.

Published comments on your own videos. Once a day we read the public comments on your recent longer videos, only those in your channel's own upload list. Google's comments service doesn't accept read-only sign-in access, so we read these public comments with a key of our own rather than asking you for a wider permission. We ask YouTube only for each comment's text, its number of likes and replies, and when it was posted: we never receive who wrote it (no name, channel or photo). We show you the comments that mention a time in your video, at that time and beside the moments viewers left or rewatched. No AI model reads them. The text is replaced at each daily read and deleted 30 days after it was read at the latest; we keep nothing calculated from comments.

We do not request access to revenue or monetary reports. We can't upload, edit, delete or comment on anything. We don't collect viewers' identities, including who wrote a comment.

What we store: your channel's ID, name, country (as you set it on YouTube) and upload list; for each video, its category and audio language as YouTube reports them; for each recent video, snapshots of its title, a fingerprint of its description (not the description itself), its thumbnail image and fingerprint, and its privacy status; hourly public counts while a video is new; and the reach report files YouTube produces.

Why: to build a record of what actually happened on your channel. That record grounds simulations you run, provides baselines, and lets us score predictions against real outcomes. It's used to provide Thommidhi to you. It isn't sold, isn't used for advertising, and isn't used to train general-purpose AI models.

Today, connected YouTube data is not sent to AI model providers. If that changes, we will update this policy before it happens.

That includes an AI assistant you connect with Oli by Thommidhi: it never receives connected YouTube data or anything we calculate from it.

Deleting the YouTube data we hold: open Creators in the app and choose Disconnect. That revokes our access with Google and deletes everything collected through the connection straight away: your channel record and upload list, titles, thumbnails, public counts, Analytics statistics, reach report files, cached insights and comments read for your videos, together with the figures we calculated from them (first-week views, your channel baseline, the “× your usual” ratio and the score built from them). You can also ask us to delete this data, or your whole account, by writing to hello@thommidhi.com; we do that within 30 days.

Revoking our access from Google: at any time, remove Thommidhi from the third-party connections in your Google Account at myaccount.google.com/connections (https://myaccount.google.com/connections?filters=3,4). We notice at the next hourly check and delete all data collected through the connection within 7 days. Disconnecting in Thommidhi deletes it immediately, so do that first where you can.

Thommidhi's use and transfer of information received from Google APIs adheres to the Google API Services User Data Policy (https://developers.google.com/terms/api-services-user-data-policy), including the Limited Use requirements.

By connecting YouTube you also use YouTube's API Services, which are subject to the YouTube Terms of Service (https://www.youtube.com/t/terms) and the Google Privacy Policy (https://policies.google.com/privacy).

Connected Instagram data

Connecting Instagram is optional and is in Beta. You connect your own Instagram professional account (Creator or Business) through Instagram's own login screen, which shows what is being requested. We ask for two permissions and use them only to read: your username, account type and recent posts, posts you're tagged in, and the comments on your own posts (which include @mentions). We never post, reply, hide or delete anything.

Instagram data is shown only to you, in Buzz. It's never shown to brands, to other creators or to anyone else, never used in Discover, the marketplace or Oli, never used for advertising, never sold, and never sent to an AI model: the requests and questions we pick out are found by fixed rules.

Text from Instagram is kept at most 30 days. Everything from Instagram is deleted at once when you disconnect, when you delete your account, or when you remove Thommidhi in Instagram's settings.

To delete your Instagram data from Thommidhi: open Buzz and choose Disconnect Instagram, or remove Thommidhi in Instagram → Settings → Website permissions → Apps and websites, or write to hello@thommidhi.com. Each deletes everything we hold from your Instagram at once. If you remove Thommidhi in Instagram, Instagram tells us and we delete it automatically.

Your use of Instagram remains subject to Meta's terms (https://help.instagram.com/581066165581870) and Instagram's privacy policy (https://privacycenter.instagram.com/policy).

How we use information

To run the service: sign-in, running simulations, storing results, and scoring predictions against outcomes you record or authorise us to collect.

To keep predictions honest: a prediction is frozen when it is made and is not edited afterwards, so later evaluation can't be shaped by hindsight.

To improve Thommidhi: we may keep aggregate, de-identified statistics, such as how often a type of prediction was right, or how many channels have connected and their combined subscriber and video counts. When you disconnect, your channel's counts stay only inside those totals, where they can't be linked back to you or your channel.

To communicate with you about your account, applications or requests, and to keep the service secure and prevent abuse.

To tell you by email when something happens in your Creator Lab: ideas written for you, a tested video's first week, a milestone in your own numbers, or news mentioning you. At most one a day; every email has a one-click link to stop them, and you can turn them off on your You page.

To say who uses Thommidhi, only with your written permission: we may name you or your channel on our site or in our materials as a creator who uses Thommidhi. We ask you first, each time, in writing; we show nothing about you beyond what you agree to; and you can withdraw that permission at any time by writing to hello@thommidhi.com.

Who processes information for us

Cloudflare hosts the application, database and email delivery. Data may be processed in data centres outside India.

AI model providers generate the parts of a simulation that use language models. They include Cloudflare Workers AI, TypeSafe AI (the Jev model) and Anthropic (Claude), and another provider only if we configure one. They receive the content of the simulation: the decision, the options, the questions, the audience description and synthetic profiles. They do not receive your password, and we don't send your email address. Each processes data under its own terms.

If you use voice input, the recording is transcribed by Sarvam AI (India), which is built for Indian languages and mixed speech, under its own terms; if Sarvam is unavailable, by a speech-to-text model (OpenAI's open-source Whisper) running on Cloudflare Workers AI, within the same Cloudflare account that hosts Thommidhi. The text of what you said, or a description you typed in its place, is then sent to one of the AI model providers above, only to sort your own words into the fields on the screen. We check the result against what you said and discard anything you didn't say. Nothing is saved until you confirm it.

To suggest an angle on an opportunity, an AI model provider receives the kind of channel you chose, the topics you typed and who you said it is for, together with a public news item. To decide which public stories fit you, TypeSafe AI's Jev model receives the kind of channel you chose, the topics you typed (and how we read them) and the languages you ticked, together with the story and the angles drafted for you, to judge whether the story fits you and whether each angle is specific, suits you and sticks to the story; its answers are kept with the opportunity. Neither receives your name, your channel's name or any connected YouTube data.

To show where you are mentioned in the news ("Your mentions"), the names you enter for it are sent as a search to the GDELT Project's public news search, and to Brave Search if we switch it on. An AI model provider then receives those names and the public headlines found, only to write a short summary that cites them. TypeSafe AI's Jev model receives those names, each public headline found through GDELT and its site, and the kind of channel and topics you gave us, only to judge whether the article is about you or someone with the same name, and whether the headline reads positive or critical of you; its answer is kept with the article, and what you mark always overrides it. Your channel's name is shown to you as a suggestion but is searched only if you save it yourself; no connected YouTube data is sent.

To understand what a brand on the marketplace is looking for, an AI model provider receives only the words the brand wrote on its page: its name, what it's promoting, what it'd like made, the kinds of brand and the languages it ticked. It answers with topics, languages, who the brand wants on camera and who it's talking to, each from a fixed list and quoting the brand's own words; if no model answers, fixed word rules do the same. The brand sees what we read, with the model's name and the date, and can correct it. No creator's information, listing, Lab data or YouTube data is sent.

To show what is happening in your topics ("Make next"), Brave Search (Brave Software) receives a question built only from the kind of channel, topics and languages you confirmed as describing you. It returns an answer written by its AI with links to its sources, which we show you and do not keep. It never receives your name, your channel's name or any connected YouTube data.

To understand topics you type in an Indian language, including in English letters ("mana biryani"), Sarvam AI (India) receives each topic you typed or said yourself. It tells us which language it is and gives it back in that language's own script and in English. Sarvam never receives topics we suggested from your uploads, your name, your channel's name or any connected YouTube data.

If you add a video file to see what happened where viewers left, Sarvam AI (India) receives only the short clips of its sound, to turn them into words with timestamps, and TypeSafe AI's Jev model receives only the words heard in each part of a clip, beside the words of your video's own opening, to judge whether that part is a sponsor read, an intro or a request to like and subscribe, or about something the opening didn't promise. Neither receives the video's title, the times of those moments, any retention or other connected YouTube data, your name or your channel's name. Our own code chooses which seconds to clip and lines the results up with your Analytics.

Google, when you sign in with Google or connect YouTube: the requests go to Google using the access you granted.

We don't sell personal information. We share it only as described here, or when required by law.

Using Thommidhi in your own AI assistant (Oli by Thommidhi, preview)

Oli lets an AI assistant you already use (for example Claude, ChatGPT or Gemini CLI) work with parts of your Thommidhi account: your ideas, brand offers and work plans. Nothing is connected unless you connect it, and you choose what each assistant may do.

What an assistant receives: only what an Oli tool returns when you or the assistant use it. That can be your creator profile as you typed it (the kind of channel you picked, the languages you ticked, the topics you typed and who you said your videos are for), your terms for brand work as you entered them in Collabs, your deal desk and saved plans, and paid briefs open to you (the brand's name and its brief). Your minimum fee is included only if you switch on sharing it; it is off unless you do.

What an assistant never receives through Oli: connected YouTube data (your videos, numbers, comments, analytics, or figures we calculate from them), other creators' information, your password, your email address or your phone number.

Your assistant's company (for example Anthropic, OpenAI or Google) receives what its tools return and handles it under its own terms and privacy policy, not ours. We don't control what it keeps or how it uses it, so check its settings, for example whether your chats may be used to improve its models.

What an assistant can write: if you allow it, it can put an offer on your deal desk and save a plan you agreed. Each is saved as coming from that assistant, and you can change or delete it.

Your control: you choose what each assistant may do when you connect it, and can change that or disconnect it at any time in your Lab (Oli). Disconnecting stops it at once: its access is revoked straight away. You can also tell us to remove a connection by writing to hello@thommidhi.com.

Copy and paste instead of connecting: you can copy a short pack of your own information, choosing what goes in it, and paste it into any chat yourself. We send nothing anywhere for this. If you paste a plan back, we read it with fixed rules and save it to your Lab.

How it connects: your assistant signs in with an industry-standard method (OAuth 2.1). It never sees your Thommidhi password. Its access passes expire after an hour and are renewed for up to 30 days while you stay connected; we store them only as one-way hashes.

Collaborations between creators and brands (Beta, limited pilot)

A person at Thommidhi reads every brand brief, and confirms its details with the brand before anything else happens.

If a creator has said they're open to briefs, that person reads the creator's collaboration profile (what the creator typed in Collabs, described above) to decide whether to offer them a brief, and writes the reasons and what's unknown. Candidates are narrowed by fixed rules on what both sides declared, such as languages, products the creator won't promote, budget against their minimum fee, and availability. No connected YouTube data, no YouTube Analytics and nothing calculated from them is used to choose or rank creators, and there are no scores.

A brief, with the brand's name, is shown to the creators we offer it to, so they can decide whether they're interested. Those creators may use their own AI assistant, including one connected with Oli, to understand the brief and plan their work. Thommidhi doesn't send briefs to AI providers itself.

A brand sees a creator only after that creator says they're interested, and then only the fields the creator chose to share: their style in their own words, what fits, the categories they welcome, languages, deliverables, example links and, only if they chose it, the rates from their media kit. A brand never sees a creator's YouTube Analytics or other connected YouTube data, their minimum fee, their Lab readings or anything else from their Lab.

What happens next (interest, a pass and its reason, an agreed brief, a completed collaboration) is recorded with dates, and outcomes are labelled with who reported them. Nothing about a collaboration is emailed to a creator or a brand automatically; introductions are made by a person.

A creator can stop being offered briefs at any time, change what a brand may see, or delete anything on their desk.

Other creators: if you switch on “Let creators find me”, creators who also switched it on see your name, channel link, kind of channel, size as you told us, and, if you chose to share them, your style and languages. Never your rates, ideas, minimum fee or anything from YouTube. Contact details in a collab suggestion are shown only after both creators say yes.

Your page for brands: if you switch it on, anyone with its link sees what you chose to share and the ideas you chose to show, and can send you a brief through Thommidhi. A person confirms every brief before you see it. Switch the page off, or ask for a new link, and the old link stops working at once.

Being found in search, only if you choose it: switching on “Let people find my page on Google and AI search” makes your page for brands public to search engines and AI assistants, which may index, show and quote it. It then carries your name, what you chose to share and, if you claimed a listing, that listing's facts (the languages you make videos in and what you make, and who appears in your videos and where you're based only if you said them yourself), each with its source and date. Never your contact details, your minimum fee or anything from YouTube. It's off unless you switch it on. Switch it off, switch the page off or ask for a new link, and it leaves our pages and sitemap at once; search engines drop it when they next crawl it, which we don't control.

The private marketplace: if you switch on “Show me to invited brands”, brands Thommidhi lets in one by one, each signing in with a one-time link sent to their work email, see the same fields as on your page. A brand can say it's interested; you see its name and note in your Lab and decide. If you both say yes, a person introduces you. Revoking a brand's access ends its sessions at once.

Open calls: a brand Thommidhi invited can describe what it's looking for. Creators whose languages and the kinds of brands they welcome fit see the brand's name and what it wrote, never its contact details. If you raise your hand, that brand (only) sees your card, with only the fields you chose to share; you can take your hand back at any time.

Local businesses we list from public sources

To help creators find local work, we list some local businesses that already promote themselves in public, for example with ads that appear in public ad libraries, a paid partnership with a creator, a sponsored listing or a launch. These businesses haven't joined Thommidhi, and a listing never says or implies that they work with us.

What we keep: the business's name, kind, city and area, a link to its own public page (Instagram, Google Maps or its website), and the public evidence that it promotes itself, with its date and a link to the source. We never keep a listed business's phone numbers or email addresses, logos or ad images.

Who sees it: only signed-in creators with a connected YouTube channel, in Collabs, with a label saying the business hasn't joined. Listings aren't on our public pages. Each business also has its own page at a private link, which we may send it, showing what we saw and its choices.

If a creator is interested: the creator raises a hand, and a person at Thommidhi may contact the business through its public contact details to offer an introduction. Contact details we use for that are kept with that introduction only.

The business's choice: it can claim its listing (which joins our local spotlight, as described above) or remove it in one click from its own page, or by writing to hello@thommidhi.com. A removed listing is hidden from creators at once and doesn't come back on its own.

How long we keep it

Account and simulation data: while your account is active. If you ask us to delete your account, we delete it within 30 days.

Connected YouTube data: statistics (views, watch time, reach) are kept while the channel stays connected. Video titles, descriptions and thumbnails are kept for at most 30 days and refreshed while connected. Figures such as "× your usual" are calculated by Thommidhi from that data and are labelled as ours, not YouTube's.

Connected YouTube data: If you disconnect in Thommidhi, we revoke our access with Google and delete the data straight away. That includes figures we calculated from it — a video's first-week views, your channel baseline and the “× your usual” ratio, along with the score built from them. If you revoke access from your Google account instead, we notice at the next hourly check and delete the data within 7 days.

What you type for your media kit: while your account is active, and you can change or clear it at any time. The kit's YouTube figures are not stored with it; they are read from your channel's current insights each time you open it.

Comments on your videos: the text is kept at most 30 days after it was read (it is replaced at each daily read while your channel is connected) and deleted immediately when you disconnect. Who wrote a comment is never received.

Connected Instagram data: text at most 30 days; everything deleted at once when you disconnect, delete your account, or remove Thommidhi in Instagram's settings.

Video files you add: never uploaded. The sound clips cut from them and the words heard in them are not kept; they exist only while the request that reads them runs. What was found at each moment is kept with that video while your channel is connected, and deleted when you disconnect.

Voice recordings and their transcripts: not kept. They exist only for the few seconds it takes to turn your words into the answers on your screen.

What you tell us about your channel, and your responses to opportunities: while your account is active. You can change your description at any time.

How we read the topics you typed (their language, and the same words in its own script and in English): while the topic stays on your profile. Removing a topic removes how we read it.

The names you ask us to look for in the news, the public articles found for them and what you mark as about you or not: while your account is active. Changing the names replaces what was found. Brave Search results are never stored.

What you type in Collabs, and your deal desk: while your account is active. You can change or delete it at any time, and it is deleted with your account.

Oli connections, the log of which tool ran and when, and the plans you saved: while your account is active, and deleted with it. Disconnecting an assistant revokes its access at once.

How you appear beyond your Lab, your collab suggestions and the interest brands showed in you: while your account is active, and deleted with it. A brand's marketplace access and its record: until the brand asks us to remove it or the pilot ends, and no more than 12 months after its last sign-in.

Brand briefs, their status and the record of what happened: while the pilot needs them, and no more than 12 months after a brief is closed, unless the brand asks us to delete them sooner. If a creator deletes their account, they are removed from every brief.

Predictions stay with the simulation they belong to, including after it is archived. Deleting your account deletes them.

Job applications: for as long as the hiring process needs, and no more than 12 months after a decision, unless you ask us to keep them longer.

Early-access requests: until we've replied and for up to 12 months after, unless you become a user or ask us to delete them sooner.

Audience poll answers: while the creator's account is active. When their account is deleted, the answers go with it.

Local businesses: one that joined, while it takes part, and deleted when it asks. One we listed from public sources, while it is listed; if it removes itself, we keep the hidden record only so that a later search never lists it again, and we delete that too if it asks.

Security

Connections to Thommidhi use HTTPS. Passwords are hashed. Platform access tokens are encrypted before they're stored, and the key is held separately from the database. Access to production systems is limited to the people who run the service.

No system is perfectly secure. If we become aware of a breach that affects your information, we will tell you without undue delay.

Your choices

Access, correct or delete your information, or get a copy of it, by writing to hello@thommidhi.com.

Use voice input or not, as you prefer. Everything it fills in can be typed or tapped instead, and you can refuse or withdraw microphone permission in your browser at any time.

Disconnect a YouTube channel at any time from Creators in the app. You can also remove Thommidhi from the third-party connections in your Google Account (https://myaccount.google.com/connections?filters=3,4); we then delete the data within 7 days.

Share an experiment only if you choose to. A public experiment link shows the titles you tested, your pick, the frozen simulation result, how it was made and, once scored, the result relative to your usual views. It never shows your channel, account or view counts, and you can turn it off at any time.

In Collabs, choose field by field what a brand may see, turn briefs from brands on or off, and say no to any brief, with or without a reason. Switch off being found by creators, being shown to invited brands, being found in search, or your page for brands, at any time.

Archive simulations in the app. Archived simulations are hidden from your workspace; to have them permanently deleted, write to hello@thommidhi.com.

Decide whether we may name you or your channel as a creator who uses Thommidhi. We never do so without asking you first; you can say no, or withdraw a yes later, by writing to hello@thommidhi.com.

Children

Thommidhi is intended for people aged 18 and over. We don't knowingly collect information from children. If you think a child has given us information, contact us and we'll delete it.

Changes to this policy

We'll update this page when what we collect or how we use it changes, and change the date at the top. If a change is significant, we'll tell account holders by email before it takes effect.

Contact

Questions or requests: hello@thommidhi.com.