Privacy Policy
Effective: August 12, 2026 · Previous versions
- We store your account, your ideas, specs, architectures, and SOWs — that's the product.
- We never store source code you submit for verification. Verification runs in memory; only the report is kept.
- Your content is processed by our AI providers (Groq and Anthropic) to generate results. Both are contractually prohibited from training on it.
- No advertising and no selling of data, ever. Only strictly necessary session cookies. One page — /founding — uses analytics, and only before sign-up.
- You can delete your account at any time; deletion is immediate and cascading. Founding members: the founder number is retained so it is never reissued — email us to erase the rest.
1. Who we are
PostIdea (postidea.app) is operated by Syed Muhammad Sufyan, an individual proprietor based in Karachi, Pakistan ("we", "us", "our"). This policy explains what information we collect when you use PostIdea, why we collect it, and the choices you have. Questions or requests: support@postidea.app.
2. What we collect
Account data
Your email address, a hash of your password (if you register with one — we never store the password itself), your OAuth provider identifier and display name (if you sign in with Google or GitHub), and your role selection from onboarding. If you join the waiting list, we store the email address and role you submit so we can contact you about access.
Content you create
Your idea conversations, specifications, architecture documents, decisions, and statements of work. This is the substance of the product and is stored in our database for as long as your account and projects exist. Your conversation and specification content is sent to one of our AI providers to generate results (see Subprocessors).
Content you submit for verification
We do not store your source code. Verification runs entirely in memory: the file listing, file contents, and code diff you submit are processed to produce a report, and then discarded. What we keep is the report — verdicts, coverage scores, and SHA-256 fingerprints of the inputs and outputs — never the code itself. The optional AI audit (off by default) sends your code diff — never full file contents — to Groq for one-time advisory review; the prompt is not stored, only its fingerprint. The temporary cache that supports safe retries holds a copy of the report with all AI-generated advisory text removed, for at most one hour.
Technical data
Standard server logs (IP address, user agent, request identifiers), security event logs (for example, failed sign-in attempts, recorded with the email and IP involved), and short-lived security counters (sign-in failure counts, rate-limit state) that expire automatically within minutes. We use these to keep the service running and to detect abuse.
Founding member programme
If you join the Founding 100 programme, we create a founding member record. It holds the email address on your account, a sequential founder number, and whether your trial has been used. Everything below is additional to that, and all of it is optional.
- Phone number, WhatsApp preference, and best time to call — collected only if you choose to fill in the welcome form after checkout. We use them for one thing: arranging your 30-minute setup call. Leaving the form blank changes nothing about your account or your access.
- “What are you building” — free text you may optionally provide, used as context for that same call.
- Channel attribution (
utm_source,utm_medium,utm_campaign,utm_content) — where you arrived from, taken from the link you clicked. Used to understand which channels reach people. Internal only. - Referral — if you arrive through someone's invite link, we store that referral code so we know who referred you. Internal only.
- Call notes — after a setup call, we may record whether the call happened and short notes about it, so we can follow up properly.
Who can see this. Your founding member record is readable by Sufyan and by two people on the team who help run onboarding and support. Nobody else. Your founder number is not currently displayed publicly anywhere — there is no public founders list or profile page.
Founding member feedback
If you send feedback from the founding member page, we store what you wrote: an optional 1–5 rating, and three optional free-text answers — what confused you, what is missing, and what would stop you paying. Why: to decide what to build and fix next; the whole reason the programme is capped at 100 people is that feedback at that size can actually be read and acted on. Who: Sufyan and the same two people on the team who handle onboarding and support. How long: kept for 24 months from submission, or until you ask us to delete it. It is never sold, never shared with advertisers, and never used for anything other than improving the product.
Write only what you are comfortable storing. These are free-text boxes, so please do not paste passwords, API keys, or client information into them — we cannot filter what a text box receives.
Landing page visit counts
When the /founding page loads we increment a counter for that day and the campaign tag in the link you followed. This is a tally, not a record of you. We store no IP address, no browser fingerprint, no session identifier and no cookie against it — the table can answer “how many people arrived from this link today” and is structurally incapable of answering “who”. We use it to see which posts bring people in.
We do not sell this data, and we never share it with advertisers. It is not used for advertising, profiling, or any purpose other than running the founding programme and supporting you.
3. AI processing
PostIdea's generation features (idea refinement, specifications, architecture, statements of work, risk scoring, and the optional verification audit) are powered by large language models hosted by third-party providers. We currently use two:
- Groq, Inc. — Llama-family models. Used for specification generation, architecture, statements of work, risk scoring, and the optional verification audit.
- Anthropic PBC — Claude models. Used for the multi-agent architecture council stage.
Which stage runs on which provider can change as we tune output quality. Both providers are listed as subprocessors below, and both are contractually prohibited from using your inputs or outputs to train or fine-tune models.
Groq is bound by the Groq Services Agreement §4.2 (effective October 15, 2025) — see Groq's Services Agreement and Data Processing Addendum. Anthropic's Commercial Terms of Service state that Anthropic may not train models on customer content submitted through its commercial API — see Anthropic's Commercial Terms of Service, Data Processing Addendum, and Privacy Policy.
To be precise about the limits of that promise: "not used for training" does not mean "never accessed." A provider may access data as necessary to operate its service, monitor for abuse, and enforce its usage policy, and may retain inputs and outputs for a limited period for those purposes under the terms of its data processing addendum.
4. Subprocessors
We use a small number of service providers to run PostIdea:
| Provider | Purpose | Location |
|---|---|---|
| Groq, Inc. | AI inference (specification, architecture, SOW, risk scoring, and the optional verification audit) | United States |
| Anthropic PBC | AI inference (architecture council stage) | United States |
| DigitalOcean, LLC | Application and database hosting | India (Bangalore) |
| Creem | Merchant of record — payment and subscription processing for paid plans | United States / EU |
| Google LLC | "Sign in with Google" authentication; web fonts on marketing pages | United States |
| GitHub, Inc. | "Sign in with GitHub" authentication; fetching file listings of public repositories you point us at | United States |
| Resend, Inc. | Transactional email — account verification codes, password-reset codes, account notices, and the founding member welcome email | United States |
| PostHog, Inc. | Product analytics on the /founding page only — anonymous page-view, scroll, and button-click events before sign-up | United States |
When you buy a paid plan, payment is handled by Creem as merchant of record. Creem processes your checkout and payment details as an independent controller under its own privacy notice — we never receive or store your full card details.
We may substitute subprocessors with providers offering equivalent data protections and will update this list when we do. We also plan to add product analytics (for example, PostHog) and error monitoring (for example, Sentry); each will be added to this list and take effect only after this policy is updated to name it. We do not use either of them today.
PostIdea's use and transfer of information received from Google APIs adheres to the Google API Services User Data Policy, including the Limited Use requirements. We access only your basic profile (email address and name) for authentication.
5. Retention
- Projects, conversations, architectures, SOWs: kept while your account and the project exist.
- Specifications: we keep the 10 most recent versions per project; older versions are pruned automatically.
- Deleted projects: held for a roughly 30-second undo window, then permanently deleted along with their specs, architectures, and conversations.
- Verification: submitted code is never stored. The report summary is kept with your project; the retry cache holds a stripped report copy for at most 1 hour.
- Risk-score reports: expire automatically after 7 days.
- Security counters: sign-in failure and rate-limit counters expire automatically within about 15 minutes.
- Founding member data: contact details you gave for the setup call (phone, WhatsApp preference, call time, what you are building) and our call notes are kept for 24 months from your last activity, or until you ask us to delete them — whichever comes first. Channel attribution and referral data are kept on the same 24-month basis.
- Founding feedback: kept for 24 months from submission, or until you ask us to delete it. Deleting your account does not automatically remove it, because it is stored against the founding record rather than the account — email us and we will erase it.
- Landing page visit counts: aggregate daily tallies per campaign tag, with nothing identifying attached. Kept for 24 months as a running record of which channels worked.
- Account deletion: removes your projects, specifications, architectures, conversations, and account record immediately and cascadingly. One deliberate exception: if you were a founding member, the founding record is kept with your account link removed, because the founder number must not be reissued to someone else and we need to know a trial was already used. To have the remaining contact details and notes erased as well, email support@postidea.app and we will delete them.
6. Cookies and third-party requests
We use only strictly necessary cookies: a session cookie that keeps you signed in (HttpOnly, so scripts cannot read it) and a CSRF-protection cookie. There are no advertising cookies and no ad-network tracking of any kind.
One analytics exception, and it is limited to a single page. The /founding landing page loads PostHog to measure how that page performs: it records that the page was viewed, which sections were scrolled to, and whether the “Claim your spot” button was clicked or led to checkout. Those events carry only the channel attribution described in section 2 — never your email, name, phone number, account id, or founder number. Nothing is measured after sign-up, and the rest of the site and the app load no analytics at all. If you block PostHog, the page and the checkout button work exactly the same.
Attribution values themselves are stored in your browser's sessionStorage for the duration of that browsing session. They are not cookies, they are not readable by other sites, and they are discarded when you close the tab.
One honest caveat: our marketing pages (including this one) load fonts and scripts from third-party CDNs (Google Fonts, Tailwind CDN). Your browser discloses your IP address to those services when it fetches these files. The app itself (/app) is stricter — it loads no third-party scripts at all, enforced by our Content Security Policy.
7. Your rights
You can access and correct your data in the app. You can copy or download your specifications and SOWs from the app directly; for a full export of your data, email us. You can delete your account and all associated data at any time from the app; deletion is immediate and cascading, with the single founding-member exception described in Retention. To have founding contact details and call notes erased, or for any other request — access, correction, deletion, portability, or objection — email support@postidea.app and we will respond within 30 days.
8. International transfers
We operate from Pakistan and use subprocessors in the locations listed above (primarily the United States). By using PostIdea you understand that your data is processed in those locations under the protections described in this policy and our subprocessors' agreements.
9. Children
PostIdea is not directed to children and requires users to be at least 18 years old. We do not knowingly collect data from anyone under 18; if we learn we have, we will delete it.
10. Security
We take specific, verifiable measures to protect your data — including HttpOnly session cookies, CSRF protection, brute-force lockouts, TLS everywhere, and a strict Content Security Policy. The full list, including what we don't yet claim, is on our Security page.
11. Changes to this policy
When we change this policy, we will update the effective date at the top and keep prior versions available at /legal/archive/. For material changes we will notify you in the app or by email before they take effect.
12. Contact
Privacy questions and requests: support@postidea.app. Security reports: security@postidea.app (see security.txt).