Case study · Built research application
Affiliate & creator hunter
A research tool for finding potential brand partners and creators, then keeping track of what happens next. Search results become a saved working list with notes and outreach stages.
Finding a name is only the first step.
A useful shortlist needs context: does the candidate reach the right audience, do they have a public affiliate programme, and have you already contacted them? I built the research and follow-up records into the same application.
- 01Business, niche & region
- 02Search-assisted research
- 03Saved candidates & outreach stages
What I built
Research around your business
The search uses a niche, region and saved business description. Prompts ask for relevant audience and brand fit, rather than simply listing businesses in the same broad category.
Two kinds of partner search
Brand research separates existing affiliate programmes from potential partnership pitches. Creator research collects public handles, platforms, fit explanations and available contact information.
A list you can work through
Candidates are saved to the account, with notes, source information and stages such as pitched, replied or negotiating. Repeat searches update existing candidates rather than creating the same entry again.
Keep the judgement with a person
The tool returns source references and research notes. Public information and AI assessments need checking before you approach someone; the application is a research aid, not a promise of a suitable or successful partnership.
What this delivers
A place to discover candidates, review the reasons for considering them and record the progress of a conversation. It connects research to the everyday job of following up.
The technical side
The implementation behind the research and saved records.
- Research
- Gemini with Google Search grounding returns structured candidate lists and source URLs. Separate prompts handle brand and creator research.
- Server
- Node.js and Express handle authenticated searches, saved business context and updates to candidate records.
- Database
- PostgreSQL stores accounts, leads, creators, notes and statuses. Account-scoped unique keys support updating repeat results.
- Accounts
- Password hashing and database-backed sessions support sign-in. Candidate records are linked to the account that owns them.
What could I build for you?
A tailored research and follow-up tool for potential partners, suppliers, stockists or business leads. We would agree the sources, useful fields and review process, then connect the results to a workflow your team can actually use.
What would help your business?
Tell me what you're researching, producing or keeping track of. We can work out which parts need a person, which could be automated, and what a useful first version would look like.
Talk through your idea