Audience research

Build personas from your website, CRM exports, and saved business profile. No ad-platform APIs required.

Workflow

Save your product, tone, and target audience with the business profile tools
Analyze your website URL to extract headings, copy, and positioning
Import CRM segments (CSV or JSON) with industry, job title, and tags
Generate draft personas with pain points, goals, and platform preferences

What is not included

Loopmark Agent does not connect to Meta Ads, LinkedIn Campaign Manager, Google Ads, or follower analytics APIs in the OSS edition. Audience research here means website + CRM + AI personas.

Data collection & limits

This is not large-scale web scraping. Audience research uses small, user-initiated fetches to help draft marketing personas.

What the agent may fetch

Website URL (you provide): one HTTP GET of public HTML per analysis. We parse title, meta tags, headings, and body copy. Response size is capped; requests time out after a few seconds.
Social profile links on that site: optional follow-up GET of public profile pages (LinkedIn, Instagram, X, etc.) to read Open Graph / title / description when available. Many platforms return login walls; we mark those as blocked and do not bypass authentication.
CRM files (you upload): CSV or JSON you export from your own CRM. Nothing is pulled from third-party CRM APIs in the OSS edition.
Business profile (you enter): product name, tone, audience description, website URL stored locally for reuse.

What we do not do

No crawling sitemaps, search results, or competitor databases at scale
No ad-platform or social analytics APIs (followers, ad spend, lookalikes)
No circumventing rate limits, CAPTCHAs, or login walls
No resale or sharing of fetched or imported data

Operational limits (self-hosted)

Volume is intentionally low (on-demand, per user action). If you deploy for many users or geos, you are responsible for robots.txt, site terms, request politeness, and infrastructure (including proxies) where appropriate.

PII & privacy

When you self-host Loopmark Agent, you are responsible for how personal data is collected, stored, and processed. The OSS agent stores data locally under data/ (JSON files) unless you integrate your own backend.

Where PII can appear

Feature	Typical fields	Storage
CRM import	Name, email, company, job title, tags, segment	`data/` (imported rows / personas)
Funnel / leads	Name, email, company, notes, score	`data/leads.json`
Complaints	Customer name, email, message	`data/complaints.json`
Website analysis	Usually no individual PII; public marketing copy only	Transient / persona drafts
Social public fetch	Public bio/title text only when pages are accessible	Enrichment on analysis result

Third parties

OpenAI: persona and content tools may send website excerpts, CRM summaries, or messages you provide to the model API. Review OpenAI’s policies and use enterprise / zero-retention options if your compliance program requires them.
Twitter, LinkedIn, Buffer: only when you configure publishing; not used for audience scraping.

Recommendations

Do not commit data/, CRM exports, or .env to git.
Import the minimum CRM columns needed (avoid full contact dumps when anonymized segments are enough).
Honor GDPR, CCPA, and local rules for consent, retention, and deletion when processing customer or lead data.
Redact or delete JSON files when offboarding users or closing a deployment.

Questions from partners or reviewers? Point them to this page and the README section.