Privacy policy
Last updated 5 November 2026
Who we are
Compilar is a marketing data warehouse made by Little Bear Apps in Melbourne, Australia. Compilar is the new name for the tool previously called Auditor Toolkit (working title channelo); you may still see those names in older documents. Compilar is in development. Today it’s used, by invitation, to prepare reports for a small number of clients. There’s nothing to sign up for.
Little Bear Apps runs Compilar. Little Bear Apps is the business name of Nathan Schram (ABN 32 380 848 852), a sole trader. In this policy, “we” means Little Bear Apps operating Compilar.
This policy covers two things: the compilar.co website, and the client data Compilar processes on behalf of our clients.
This website
compilar.co has a few pages: the home page, the FAQ, this privacy policy, the terms, the data deletion instructions and how connecting works. It has no accounts or comment boxes, its only form is the register-interest form described below, and it runs no advertising. It is hosted on Cloudflare, which processes standard request data (such as your IP address) to deliver and protect the site. See Cloudflare’s privacy policy.
When you register interest {#register-interest}
If you register interest, we ask for your email address and, if you choose to give them, your first name and which best describes you (solo marketer, consultant or freelancer, in-house team, agency, or something else). You tick what you’d like: an email when beta applications open, occasional build updates, or both. We also record which version of the form you agreed to, and when.
You then get one email asking you to confirm your address, sent through Resend, our email provider. Its link works once, for 7 days, and until you confirm, nothing else is sent. Once you confirm, your address is added to the Compilar lists you chose.
- Your choices: every email has a link to a page where you can change what you get or unsubscribe, and your mail app’s unsubscribe button works straight away. You can also email privacy@compilar.co to be removed.
- What we use it for: only to send what you asked for. We don’t sell your details or share them with anyone other than the services named here. The emails don’t track whether you open them or which links you click.
- Spam protection: the form uses Cloudflare Turnstile to check you’re a person. Your IP address is combined with the date and turned into a code (a hash), which is kept for 10 minutes to limit repeated attempts.
- Where it’s kept: your sign-up is stored with Cloudflare, which hosts this site, and once you confirm, with Resend (stored in the United States; emails are sent from Japan).
- How long: a sign-up that’s never confirmed is deleted after 30 days. A confirmed sign-up is kept until you ask us to remove it; if you unsubscribe from everything, we keep a note that you did, so you’re not emailed again. Resend keeps copies of sent emails for 30 days.
Analytics and cookies
We use Google Analytics 4, loaded through Google Tag Manager, to count visits and see which pages get read. It only sets cookies (_ga and _ga_*) if you choose Accept in the cookie banner.
- Accept: Google Analytics sets its cookies and records page views, clicks on “Register interest” and general details such as browser, device type and approximate location (country or city).
- Decline, or no choice: no analytics cookies are stored. Google receives cookieless pings that don’t identify you, which it uses to estimate overall traffic.
We don’t use advertising features, remarketing audiences or cross-site profiles. Your choice is saved in your browser’s local storage (analyticsConsent) so the banner doesn’t keep coming back.
Client data: our role
When a business engages us, it grants us read access to its marketing platforms. That business is the controller of its data. We, operating Compilar, are the processor: we process the data only to prepare that business’s own reporting and analysis. Client data stays the client’s property.
We’ll sign a Data Processing Agreement on request, at no charge. It covers our role as a processor, the authorised purposes (analytics ingestion, reporting, archival), the sub-processor list below, security measures, breach notification, data subject requests, and deletion at the end of the engagement.
What client data Compilar ingests
Compilar reads data the client’s platforms already hold, through access the client grants and can revoke on the platform at any time.
| Platform | What is ingested |
|---|---|
| Google Analytics 4 | Aggregated session, event, conversion, e-commerce and audience metrics. No individual user journeys and no user-level identifiers. |
| Google Ads | Campaign, ad group, ad, keyword, audience and conversion data, aggregated. No payment methods and no customer-list contents (audience names only). |
| Google Search Console | Aggregated search query, page, country and device data; sitemap and indexing status. No per-user search history. |
| Google Tag Manager | Tag, trigger, variable, workspace and version metadata (the configuration of the tag setup). No event payloads. |
| Google Business Profile | Business information (name, hours, categories), performance metrics, and reviews, including reviewer display names (public data Google already shows on Maps). |
| Meta (Facebook, Instagram, Ads) | Page and Instagram Business insights, follower counts, ad campaign performance, audience definitions, aggregate Pixel/Dataset event data. No message contents, comment threads, post or caption text, or payment methods. Access is read-only, through partner access to our Meta Business Portfolio or Facebook Login for Business (see “How Meta access works” below). |
| Plausible Analytics (if used) | Aggregated visitor, page, event and source data. Plausible is cookieless, so there are no individual user identifiers to ingest. |
| Bing Webmaster Tools (if used) | Aggregated weekly search query and page performance (clicks, impressions, average position). No per-user search history and no country or device breakdown. |
Google, Meta and Microsoft (Bing) are the source platforms here, not our sub-processors: they already hold this data.
Meta Lead Ads. Where an engagement includes Lead Ads reporting, Compilar ingests lead volume metadata only: how many leads each form or ad received per day, when the first and last lead of the day arrived, and the form’s name and status. The answers a person typed into a lead form (name, email, phone or anything else) are discarded at the point of retrieval, never stored and never logged. A lead’s identity can’t be reconstructed from the warehouse.
Competitor benchmarks (optional). Where a client nominates competitor Instagram accounts or Facebook Pages, Compilar reads the same public counts anyone can see on those profiles (follower, following and post counts, and per-post like, comment and view counts) through Meta’s official endpoints. No captions, comments or messages are requested or stored, and these results are for the operator only.
Search Intelligence. Compilar also queries search-data providers (mainly DataForSEO) for search rankings, keyword volumes and competitive analysis on the client’s domain. This reads public search results, not the client’s platform data, and needs no access grant from the client.
How Meta access works. A client gives Compilar access in one of two ways: it adds our Meta Business Portfolio as a partner on the assets it chooses (Facebook Page, Instagram Business account, ad account, Pixel or Dataset), or it connects through Facebook Login for Business. No passwords are shared. We ask for the minimum permissions reporting needs: pages_show_list, pages_read_engagement, read_insights, instagram_basic, instagram_manage_insights, business_management and ads_read. Compilar only reads through this access. A client can remove our partner access in its Meta Business Settings at any time (or, if it connected through Facebook Login for Business, remove Compilar under Business integrations in its Facebook settings), which cuts off our access to that client’s assets straight away without affecting any other client. Meta’s own handling of this data is covered by Meta’s privacy policy. The contract side (authorisation, revocation and the Meta rules we work under) is in our terms.
If you’re a Facebook or Instagram user and your data appears in a client’s Meta accounts, you can ask us directly to delete anything we hold about you. See our data deletion instructions.
LinkedIn (not in use yet). Compilar’s LinkedIn connection isn’t live, so we don’t hold any LinkedIn data. When it is, a client will give us a role on its LinkedIn ad account and, where relevant, its Company Page, and Compilar will read campaign performance and Company Page analytics for that client’s reports. It will only read. We’ll update this policy before the connection is used for any client. See LinkedIn’s privacy policy.
Google user data
When a client connects Google to Compilar, Google issues us an access token for the products the client chose. This section covers the data we receive that way.
- What Compilar reads, for the products the client ticks:
- Google Analytics: reports and property settings (
analytics.readonly). - Search Console: search performance and site status (
webmasters.readonly). - Tag Manager: container configuration (
tagmanager.readonly). - Google Ads: campaign performance (
adwords). Google has no read-only Ads scope; we only read. - Business Profile: performance metrics, reviews and listing details (
business.manage; read-only in practice). - The email address of the Google account that connects, so we know which account granted access.
- Google Analytics: reports and property settings (
- How we use it: only to prepare that client’s own reports and analysis.
- Who reads it: Nathan Schram, to prepare that client’s reports. The client agrees to this on the connect page, before Google’s consent screen. Nobody else reads it, apart from the service providers named in this policy that host or process it for us.
- Storage and security: Google’s tokens are encrypted before they’re stored on our server in Germany, with the encryption key held separately in our secrets manager (Bitwarden Secrets Manager). The data goes into that client’s own warehouse file.
- What we never do: sell it, use it for advertising, build profiles of individuals from it, combine it across clients, or use it to develop, improve or train AI or machine-learning models. The only AI services that receive it are Anthropic and OpenAI, for analysis (see AI services and your platform data). Beyond that, we share it only with our hosting provider (Hetzner) or where the law requires it.
- Disconnecting: remove Compilar at myaccount.google.com/permissions, or email privacy@compilar.co. We then revoke the token and delete our stored copy within 5 business days. Warehoused data is deleted as set out in our data deletion instructions.
- Limited Use: Compilar’s use and transfer of information received from Google APIs to any other app will adhere to the Google API Services User Data Policy, including the Limited Use requirements.
Where the data lives
| What | Where |
|---|---|
| Processing and the warehouse | A single Hetzner Online GmbH server in Falkenstein, Germany. Each client’s data is in its own separate database file. |
| Backups | Snapshots in Hetzner Object Storage (Falkenstein, Germany). |
| Raw API responses (the platform’s own reply, kept so a processing fault can be corrected from the original) | The same Hetzner Object Storage, under the client’s own prefix. Operator-only; never shown in a report or the client portal. |
| Credentials (API keys, access tokens) | Access tokens a client grants through the connect flow are encrypted before they’re stored on our server in Germany; the encryption key is held in Bitwarden Secrets Manager (US-hosted). API keys and our own credentials are held in Bitwarden Secrets Manager and retrieved only at run time, apart from one Google service-account key file stored on the server, readable only by the service and never in code. |
| Logs | Structured logs on the same server. A scrubbed copy of operational logs, metrics, traces, performance profiles and run records (job, test and data-quality outcomes as ids, counts and hashes), plus browser-session and synthetic-check results from our own operator dashboard and from the client portal’s Google connect pages (page-load timing, Web Vitals and script errors, the browser type and the network address it connects from — never a name, email or account id, and every web address is cut before its ?, so a connect link or Google sign-in code never leaves your browser), goes to Grafana Labs (Grafana Cloud, EU region: AWS Frankfurt, Germany). It carries no analytics data and no client identifier beyond the client’s internal profile id. Since 12 October 2026, server and job logs record no IP addresses or other personal information. |
Sub-processors
These are the only third parties that receive client data, grouped by what reaches them. We’ll tell clients in writing before a new one is added.
Infrastructure
- Hetzner Online GmbH (Falkenstein, Germany): server hosting and Hetzner Object Storage for backups and raw API responses. Standard cloud-provider access; no read access to client data unless lawfully compelled.
- Bitwarden Inc. (US): credential vault. Encrypted at rest; Bitwarden can’t read it.
- Cloudflare, Inc. (global network): the secure network edge in front of our dashboard and the client portal. Request metadata and portal traffic in transit.
- Grafana Labs (Grafana Cloud, EU region: AWS Frankfurt, Germany): operational monitoring. Scrubbed service and job logs, metrics, traces and profiles, and browser-performance data from the client portal’s connect pages, used to detect and fix failures. No analytics data and no platform credentials; the only client reference is the internal profile id.
- Auth0 (Okta, Inc.) (Australian region): client-portal sign-in. The portal users’ name and email address.
Analysis and AI
- Anthropic (Claude) and OpenAI (ChatGPT, Codex) (US): AI-assisted analysis. Extracts of the client’s warehouse data while we analyse it. These are the only AI services that receive client platform data, and neither trains its models on it. Also used, through their APIs, for AI-search (GEO) checks.
- Google (Gemini), Perplexity and xAI (Grok) (US): AI-search (GEO) checks only, where enabled. The client’s business name, domain and search questions. They never receive client platform data.
Search data
- DataForSEO: Search Intelligence. The client’s domain, business name and keywords.
- Firecrawl, Brave Search and Google (PageSpeed Insights, Knowledge Graph Search): public page addresses and search queries.
Operations
- GitHub (Microsoft) and Trello (Atlassian): Our code, issues and task board, which may name the client’s business and operational details.
- Slack and Telegram: operational alerts and the operator’s remote work sessions, whose text may name the client’s business or include summary figures.
AI services and your platform data
- Platform data goes to two AI services only. Data received through Google APIs, from Meta’s platform or from Microsoft Bing Webmaster Tools is processed by Anthropic (Claude) and OpenAI (ChatGPT, Codex) only, as our service providers, solely to analyse that client’s data and prepare that client’s own reports. Model training is switched off on both.
- The AI-search engines receive search prompts only. For AI-search (GEO) checks, Google (Gemini), Perplexity and xAI (Grok), and Anthropic and OpenAI when used for the same checks, receive the prompts being tested: a business name, a domain and the questions. They never receive client platform data: no warehouse rows, no report content and no analysis text.
- Research and search-data tools never receive client platform data. DataForSEO, Firecrawl, Brave Search and Google’s PageSpeed Insights and Knowledge Graph Search receive public page addresses, domains, business names, keywords and search queries only.
- No AI training. Client data is never used to develop, improve or train any AI or machine-learning model, ours or anyone else’s.
- Google API data. Compilar’s use and transfer of information received from Google APIs to any other app will adhere to the Google API Services User Data Policy, including the Limited Use requirements. More detail is under Google user data.
International transfers
For Australian clients. Client data leaves Australia when it’s ingested. Processing, storage and backups are in Germany (Hetzner, Falkenstein), as is the scrubbed monitoring copy (Grafana Labs, AWS Frankfurt), which contains no analytics data. Hetzner stores the data in Germany, where the GDPR applies. We rely on the GDPR Standard Contractual Clauses where required and on the recipients’ GDPR-grade controls. Transfers to the US providers listed above rely on reasonable steps under Australian Privacy Principle 8 (their published privacy and no-training terms). Extracts of warehouse data reach Anthropic and OpenAI (US) during AI-assisted analysis. Business names, domains and search questions go to the US-based AI and search-data providers listed above when AI-search (GEO) checks or Search Intelligence are enabled.
For EU clients. The warehouse, backups and monitoring data stay in the EU. Extracts of warehouse data reach Anthropic and OpenAI (US) during AI-assisted analysis. Business names, domains and search questions go to the US-based AI and search-data providers listed above when AI-search (GEO) checks or Search Intelligence are enabled. Portal sign-in (Auth0, Australia), portal traffic (Cloudflare’s global network) and our operations tools also operate outside the EU. Bitwarden (US) holds credentials and encryption keys, not client data.
How long we keep it
| Data | Retention |
|---|---|
| Warehouse data and account configuration history | Kept for the life of the engagement so results can be compared over time and past results explained; deleted within 30 days of termination or on request. |
| Backups | Daily copies kept 7 days, weekly copies 4 weeks, monthly copies at most 90 days. |
| Raw API responses | 120 days (Google Ads change history, Bing Webmaster Tools, DataForSEO); 45 days (Google Analytics report pages); then deleted automatically. |
| Server logs | 90 days |
| Monitoring copy at Grafana Labs | At most 31 days. |
| Reports delivered to the client | The client keeps them. We keep a copy in the warehouse for the life of the engagement. |
| When an engagement ends | The client’s warehouse profile, backups and retained raw API responses are deleted within 30 days, unless the client instructs otherwise in writing. |
| On request | Deleted within 30 days of a written request. |
What we don’t do
- We don’t combine data across clients. Each client’s data is in a physically separate database file.
- We don’t use client data to train any AI or machine-learning model.
- We don’t sell, rent or share client data, or pass it to anyone other than the sub-processors listed above.
- We don’t use client data for our own marketing, prospecting or audience-building.
- Compilar doesn’t post, edit, publish, advertise, build audiences or change settings through the access a client grants. It only reads.
Security
- One operator. Nathan Schram is the only person with shell access to the server and the only holder of the master credentials.
- Credentials. Access tokens a client grants are encrypted before storage, with the key held separately in Bitwarden Secrets Manager. Our own credentials are held in Bitwarden Secrets Manager, apart from one Google service-account key file stored on the server, readable only by the service and never in code.
- Least privilege. We ask for read-only access wherever the platform offers it.
- Breach notification. Under a DPA, we notify the client within 72 hours of a confirmed incident.
Your rights and contact
If your business is a Compilar client, you can ask for a copy of everything we hold about your business, have it corrected, or have it deleted, by emailing privacy@compilar.co. We acknowledge requests within 5 business days and resolve them within 30 days.
If you’re a customer or website visitor of one of our clients, or your data appears in a client’s Google or Meta accounts, that client controls your data. Contacting them is usually fastest, and we’ll action their instructions, usually within 30 days. You can also ask us directly: see our data deletion instructions.
You also have rights under the Australian Privacy Act 1988 and, where it applies to you, the GDPR. If you’re not satisfied with our response, you can complain to the Office of the Australian Information Commissioner.
Automated decision-making
Compilar brings aggregate data from clients’ advertising and analytics platforms into a private warehouse, and produces campaign-level and aggregate reports. It doesn’t score individuals, make automated decisions about specific people or take automated actions that affect individuals. No automated decision-making with a legal or similarly significant effect on individuals takes place in Compilar. That means no automated credit, employment, insurance or pricing decisions, no automated content moderation, and no automated exclusion of named people from audiences.
If a future version of Compilar adds analysis at the level of individual people, we’ll update this policy before that feature is used for any client. This follows the automated decision-making disclosure obligations that start on 10 December 2026 under the Privacy and Other Legislation Amendment Act 2024.
Changes to this policy
We’ll update this page when what we do changes, and change the “Last updated” date above. Clients are told in writing before a new sub-processor is added.