YouTube Partnership Channel Scraper — every-3-days discovery of partnership channels
Every 3 days, find YouTube channels making content adjacent to your product — filter by audience size + country + ICP relevance, fetch contact emails, write to a deduped partnership table. Pair with the YouTube Outreach Writer to actually pitch them.
Build this with agnt_
Skip the copy-paste. We'll spin up a builder session prepopulated with this blueprint's spec — providers, schedule, database schema, and the questions the agent should ask you to personalize it for your product.
Sign up free · no credit card
The motion
YouTube creators with audiences that overlap your ICP are partnership goldmines — sponsorships, integrations, mutual promotion, affiliate. But finding them by hand is brutal: searching YouTube, sifting through hundreds of channels per topic, manually checking which ones have contact emails. This agent automates the discovery. Every 3 days it runs a TypeScript skill that searches YouTube for your topic phrases, dedupes against what's already discovered, fetches subscriber counts + country + recent videos for each new candidate, applies cheap filters first (audience size, country), then LLM-judges ICP relevance from the channel's actual recent video titles. Survivors get email-fetched; survivors-with-emails get a one-paragraph channel summary written by the LLM and saved with a tier classification (small / mid / large). Per-row commits — if the skill crashes mid-run, what's saved stays saved. Pair with `youtube-partnership-outreach-writer` for the activation half.
Each channel gets tagged small / mid / large based on subscriber count. The outreach writer uses the tier to pick a different voice + offer per group.
The LLM relevance gate runs on recent video titles + channel description — what the channel actually makes — not on whether they showed up in a keyword search.
Each new channel is committed as soon as enrichment finishes. A mid-run crash doesn't lose the channels that already landed.
Channels without a discoverable email are dropped — unreachable channels never clutter the table.
Every 3 days, the agent searches YouTube for your topics, dedupes, fetches subs + country + recent videos for each new candidate, LLM-judges ICP relevance on the actual channel content, fetches emails for survivors, and writes everything to a partnership table. Per-row commits keep the run resilient.
Build this with agnt_
Skip the copy-paste. We'll spin up a builder session prepopulated with this blueprint's spec — providers, schedule, database schema, and the questions the agent should ask you to personalize it for your product.
Sign up free · no credit card
Or copy a prompt into another platform
Prefer to build with OpenClaw, Hermes, or Claude Code? Drop this prompt into your agent of choice — it seeds the goal, the agntdata endpoints to use, and a step-by-step plan.
You are helping me build a YouTube Partnership Channel Scraper. Every 3 days, it searches YouTube for channels making content adjacent to my product, filters by activity + ICP relevance + country, fetches a contact email per channel, generates a factual one-paragraph channel summary, and saves new channels to my workspace database — ready for the YouTube Outreach Writer to pitch.
REFERENCE DOCS
- agntdata API documentation: https://agnt.mintlify.app/apis/overview
- YouTube endpoints used: `Youtube_Search`, `Channel_Details`, `Channel_Videos`, `Get_Channel_Email_By_URL`.
ABOUT MY MOTION
- Product name: <YOUR PRODUCT>
- One-line description: <WHAT IT DOES>
- ICP your YouTube partners should reach: <ICP>
- Search topics (5–10): the YouTube searches that surface channels in your space. Examples for an AI-agent product: "Claude AI", "AI agents", "GTM tools", "cold email automation", "Clay tutorials", "n8n workflows".
- Country filter: <COUNTRY_CODE> (default US — partnership conversions are localized)
- Tier thresholds: small (<SMALL_MAX> subs), mid (SMALL_MAX–MID_MAX), large (MID_MAX+). Each tier gets a different outreach voice downstream. Default: <10k / 10k–100k / 100k+.
WHAT TO BUILD
- A scheduled agent (claude-haiku-4-5 — most work is in a skill).
- A TypeScript skill called `scrape_youtube_partnership_channels` that does the heavy lifting: search → filter → enrich → save row-by-row. Per-row commits so a mid-run crash doesn't lose progress.
- Save to `youtube_partnership_channels`: channel_url (PK), channel_name, subscriber_count, country, tier, contact_email, channel_summary, recent_videos (jsonb of top 5), status default 'new'.
GUARDRAILS
- ICP relevance is the LLM's only judgment call — apply it on the channel summary + recent video titles, NOT on search keywords (search is too noisy).
- Dedup against `youtube_partnership_channels.channel_url` before any enrichment call.
- Skip channels without a discoverable contact email — they're impossible to reach.
- Per-row commits, not bulk transactions. If the skill crashes after row 12 of 30, the first 12 stay saved.
When you're ready, ask me the ABOUT MY MOTION block.Paste into OpenClaw to scaffold this agent. Tweak the inputs and goal at the top of the prompt.
How to build it
6 steps. Each one links to the underlying agntdata endpoints — open them in a new tab to inspect parameters and pricing as you build.
One key for the four YouTube tools used here.
5–10 search phrases + sub-count cutoffs for small/mid/large.
Blueprint creates it on first run. UNIQUE on channel_url.
Meta-agent asks topic terms + tier thresholds, deploys the agent + skill.
`0 6 */3 * *` default. Dry-run on one topic — verify 3 channels with summaries land before turning the schedule on.
See `youtube-partnership-outreach-writer`. Reads new rows from the same table.
Endpoints used
The agntdata endpoints this blueprint depends on. All available with one API key.
Youtube Search
/search/
Topic-based YouTube search — surfaces candidate channels by recent activity.
View endpoint docsChannel Details
/channel/details
Per-channel subscriber count, country, description.
View endpoint docsChannel Videos
/channel/videos
Top 5 recent video titles — the input to the ICP-relevance LLM judgment + channel summary.
View endpoint docsGet Channel Email by URL
/channel/email
Contact email lookup. Channels without a discoverable email are dropped.
View endpoint docsShip this blueprint today
One click spins up a builder session prefilled with this blueprint's spec. We'll ask you a handful of personalization questions, then generate the agent.
Related blueprints
Browse all →Monitor 15+ subreddits twice a day for prospects describing the exact pain your product solves. AI-scored, deduplicated, and pushed to Slack — for around $4 a month.
Turn the people quietly liking and commenting on LinkedIn posts about your space into a deduped pipeline of warm leads — refreshed every day, no manual scrolling.
Turn the people quietly replying under hot posts in your topic space on X into a deduped pipeline of warm leads — refreshed every day, no doom-scrolling required.