Skip to content
BranchesTags

Repository files navigation

steadyfetch n8n templates

Free n8n workflows for ad-creative, Instagram and keyword data pipelines. No community nodes required — plain HTTP, importable into any n8n (cloud or self-hosted).

Library-ready (Sep 2026): every workflow uses an n8n Header Auth credential for Apify (no token field), writes to Google Sheets, and opens with an overview sticky note plus section notes. Five of them are scheduled monitors: they loop over your targets, remember what they already reported with the native Remove Duplicates node, label each new row with an AI Information Extractor step (Groq by default, a drop-in swap for OpenAI or Anthropic), and send a digest to Slack, Gmail or Telegram only when there is something new.

Facebook Ad Transcripts — scrape competitor ads → hooks, CTAs & transcripts

What it does: searches the Facebook Ad Library for video ads → transcribes every creative → gives you one row per ad with the first-3-seconds hook, CTA, advertiser, and full transcript. The manual version of this pipeline is sold as a paid template elsewhere; this one is free.

Download the workflow JSON →

3-minute setup

  1. In n8n: Workflows → Import from File → pick facebook-ad-transcripts.workflow.json.
  2. Create an HTTP Header Auth credential in n8n (name Authorization, value Bearer YOUR_APIFY_TOKEN, token from Apify → Settings → Integrations) and select it on the HTTP Request nodes.
  3. (Optional) Edit adLibrarySearchUrl — any Facebook Ad Library search URL works (set media_type=video), and maxAds.
  4. Click Test workflow.

What you get per ad

field example
status transcribed (only these are charged)
advertiser Native
hook_first_3s "Look, women are always right. My girlfriend told me…"
cta Shop Now
transcript full speech-to-text, any video length
seconds 32.8
charged true / false — expired links & music-only ads are $0

The last node appends one row per item to Google Sheets (connect your Google account and pick a sheet), or swap it for Airtable, Slack, or a database.

Cost

Swap-friendly

The transcription step accepts any ad-library scraper's output (it deep-scans rows for the video URL + metadata) — if you already use a different scraper, just point its results at the Transcribe Ads node body.

Competitor ad teardown — one competitor page/keyword → hooks, CTAs & transcripts, one row per ad

What it does: point it at one competitor (their Facebook Page URL or an Ad Library search URL) → scrapes their active video ads → chains the transcript actor by dataset ID (no ad rows pass through n8n, so big batches stay light) → gives you a spreadsheet-shaped teardown: one row per ad with the first-3-seconds hook, CTA, transcript length in words, language, seconds, and the full transcript. Transcribed ads first, longest first.

Download the workflow JSON →

3-minute setup

  1. In n8n: Workflows → Import from File → pick competitor-ad-teardown.workflow.json.
  2. Create an HTTP Header Auth credential in n8n (name Authorization, value Bearer YOUR_APIFY_TOKEN, token from Apify → Settings → Integrations) and select it on the HTTP Request nodes.
  3. Set competitorUrl — a competitor's Facebook Page URL (e.g. https://www.facebook.com/nike) or any Ad Library search URL (set media_type=video) — and maxAds.
  4. Click Test workflow. The flow polls the scraper run until it finishes, then transcribes.

What you get per ad

field example
advertiser Nike
ad_id 1318059422847100
status transcribed (only these are charged)
hook_first_3s "Your sister's on the phone."
cta Shop Now
seconds 36.7
transcript_words 58
language en
charged true / false — expired links & music-only ads are $0
transcript full speech-to-text, any video length

The last node appends one row per item to Google Sheets (connect your Google account and pick a sheet), or swap it for Airtable, Slack, or a database.

Swap the scraper

The scrape step is a plain HTTP call to curious_coder/facebook-ads-library-scraper ({ "urls": [{ "url": ... }], "count": N }). Any other Ad Library scraper works too: change the actor slug and JSON body in the Start Ad Library scrape node — the transcript step reads the run's dataset ID and deep-scans rows for the video URL, whatever the field names.

TikTok Top Ads transcripts — Creative Center industry/region → hooks & transcripts

What it does: pick a Creative Center market, lookback window and industry → scrapes the Top Ads with azzouzana/tiktok-creative-center-top-ads-scraper → chains steadyfetch/tiktok-ads-transcript-scraper by dataset ID right away (no ad rows pass through n8n, and TikTok's ~6-hour video links are still fresh) → gives you a spreadsheet-shaped table: one row per ad with the first-3-seconds hook, CTR, likes, brand, ad title, transcript length in words, language, seconds and the full transcript. Transcribed ads first, highest CTR first.

Download the workflow JSON →

3-minute setup

  1. In n8n: Workflows → Import from File → pick tiktok-ad-transcripts.workflow.json.
  2. Create an HTTP Header Auth credential in n8n (name Authorization, value Bearer YOUR_APIFY_TOKEN, token from Apify → Settings → Integrations) and select it on the HTTP Request nodes.
  3. Set countryCode (e.g. US, GB, DE, SA), period (7, 30 or 180 days), industry (All industries or one Creative Center industry name) and maxAds.
  4. Click Test workflow. The flow polls the scraper run until it finishes, then transcribes by dataset ID.

What you get per ad

field example
brand Promote
ad_title "Why Promote Stands Out…"
material_id 7662340784365568017
status transcribed (only these are charged)
hook_first_3s "Looking for more ways to support your growth on TikTok?"
ctr 0.39
likes 19
seconds 44.2
transcript_words 61
language English
charged true / false — expired links and music-only ads carry no result fee
transcript full speech-to-text, any video length

The last node appends one row per item to Google Sheets (connect your Google account and pick a sheet), or swap it for Airtable, Slack, or a database. Expect a fair share of uncharged no_audio rows — many Top Ads are music plus on-screen text.

Cost

Swap the scraper

The scrape step is a plain HTTP call to azzouzana/tiktok-creative-center-top-ads-scraper ({ "countryCode", "period", "industry", "maxItems" }). Any Creative Center scraper whose rows carry the TikTok CDN video link works — doliz/tiktok-creative-center-scraper, datapeak/tiktok-creative-center, beyondops/tiktok-ad-library-scraper, lexis-solutions/tiktok-top-ads-scraper, dltik/tiktok-creative-center and others: change the actor slug and JSON body in the Scrape Creative Center Top Ads node. The transcript step reads the run's dataset ID and deep-scans rows for the video link and material ID, whatever the field names.

Freshness

Creative Center video links expire about 6 hours after the scraper minted them — this flow transcribes immediately after the scrape. Expired rows come back as uncharged unavailable_expired; if the row carries the ad's material ID, the transcript actor fetches a fresh link first.

LinkedIn competitor ad monitor — weekly hooks, AI analysis & a Slack digest

What it does: every Monday it walks your list of competitor LinkedIn Ad Library pages, scrapes each one with silva95gustavo/linkedin-ad-library-scraper, chains steadyfetch/linkedin-ads-transcript-scraper by dataset ID (no ad rows pass through n8n), drops every ad it has already reported with the native Remove Duplicates node, then has an AI classify the new ones — hook type, offer, CTA type, target persona, angle and why it works. New rows go to Google Sheets and the five longest new hooks go to Slack. A quiet week writes nothing and sends nothing.

Built for: B2B demand-generation and ABM managers watching a fixed competitor set.

Download the workflow JSON →

3-minute setup

  1. In n8n: Workflows → Import from File → pick linkedin-ad-transcripts.workflow.json.
  2. Create an HTTP Header Auth credential in n8n (name Authorization, value Bearer YOUR_APIFY_TOKEN, token from Apify → Settings → Integrations) and select it on the three HTTP Request nodes.
  3. Add a Groq credential on the Groq chat model node — or drop in an OpenAI / Anthropic chat model node in its place.
  4. Connect Google Sheets and Slack.
  5. Open Config: paste one Ad Library URL per line into competitors (https://www.linkedin.com/ad-library/search?accountOwner=hubspot&countries=US&dateOption=last-30-days), set maxAdsPerCompetitor and slackChannel.
  6. Click Test workflow, then leave the weekly Schedule Trigger on.

What you get per ad

field example
advertiser · ad_id · headline HubSpot · 1508450294 · "We are HubSpot Elite Partner"
hook_type (AI) question / statistic / pain-point / story / bold-claim / other
hook_first_3s "If your revenue team is working harder than ever, but closing less,"
offer · cta_type (AI) "free CRM audit" · demo
target_persona (AI) RevOps leader
angle_summary · why_it_works (AI) ≤20 words · ≤25 words
seconds · transcript_words · language 43.9 · 112 · English
status transcribed or image_text_extracted (the charged rows) — document, carousel, article and text-only ads show non_video_skipped
charged true / false — text-free creatives, silent videos and blocked pages carry no result fee
transcript · detail_url full speech-to-text · https://www.linkedin.com/ad-library/detail/1508450294

Rows are appended to Google Sheets; the digest goes to Slack. Swap either for Airtable, Teams or email.

Cost

  • Ad Library scraping: silva95gustavo/linkedin-ad-library-scraper — $2.00–4.00 / 1,000 ads at the time of writing (check its Pricing tab)
  • Transcription: steadyfetch/linkedin-ads-transcript-scraper — from $8.00 / 1,000 ad creative transcripts, charged only when text is delivered
  • A 25-ad test run ≈ $0.60 at most (video and image ads both deliver; text-free creatives are not charged). Apify's free $5 credit covers it many times over.

Swap the scraper

The scrape step is a plain HTTP call to silva95gustavo/linkedin-ad-library-scraper ({ "startUrls": [{ "url": ... }], "resultsLimit": N }). Any Ad Library scraper works — dz_omar/linkedin-ads-scraper, memo23/linkedin-ads-scraper, ivanvs/linkedin-ad-library-scraper, automation-lab/linkedin-ad-library-scraper and others: change the actor slug and JSON body in the Scrape this competitor's Ad Library node. The transcript step reads the run's dataset ID and deep-scans rows for the LinkedIn video link or ad detail link, whatever the field names. You can also skip the scraper entirely — the transcript actor searches the Ad Library itself when given advertiser names or keywords.

Google video ad tracker — weekly script breakdowns by email

What it does: every Monday it walks your list of advertiser domains in the Google Ads Transparency Center, scrapes each one with silva95gustavo/google-ads-scraper, chains steadyfetch/google-ads-video-transcript-scraper by dataset ID (no ad rows pass through n8n), drops every creative it has already reported with the native Remove Duplicates node, then has an AI break the new scripts down — hook type, promise, proof element, CTA type and script structure. New rows go to Google Sheets and Gmail sends the week's breakdown, counting transcribed and silent ads separately.

Built for: YouTube and Google Ads performance marketers tracking a set of brands.

Download the workflow JSON →

3-minute setup

  1. In n8n: Workflows → Import from File → pick google-ads-video-transcripts.workflow.json.
  2. Create an HTTP Header Auth credential in n8n (name Authorization, value Bearer YOUR_APIFY_TOKEN, token from Apify → Settings → Integrations) and select it on the three HTTP Request nodes.
  3. Add a Groq credential on the Groq chat model node — or drop in an OpenAI / Anthropic chat model node in its place.
  4. Connect Google Sheets and Gmail.
  5. Open Config: one advertiser domain per line in advertiserDomains (hellofresh.com), plus region, maxAdsPerDomain and digestRecipient.
  6. Click Test workflow, then leave the weekly Schedule Trigger on.

What you get per ad

field example
advertiser · creative_id HelloFresh SE · CR17857019577632817153
hook_type (AI) question / statistic / pain-point / story / bold-claim / demo / other
hook_first_3s "What if your food was smarter?"
promise (AI) "dinner solved in 15 minutes"
proof_element (AI) testimonial / demo / stat / none
cta_type · script_structure (AI) shop now · problem-solution
summary (AI) ≤25 words
seconds · transcript_words · language 15 · 34 · en
status transcribed (only these are charged)
charged true / false — removed videos, music-only ads and rows with no video carry no result fee
transcript · video_url full speech-to-text · https://www.youtube.com/watch?v=ONzIwHyQGJs

Silent ads still reach the sheet with their status, so nothing goes missing. Rows are appended to Google Sheets; the digest goes out by Gmail — swap it for Slack or Teams.

Cost

  • Transparency Center scraping: silva95gustavo/google-ads-scraper — $1.90 / 1,000 ads at the time of writing (check its Pricing tab)
  • Transcription: steadyfetch/google-ads-video-transcript-scraper — from $8.00 / 1,000 video ad transcripts, charged only when a transcript is delivered
  • A 25-ad test run ≈ $0.55 at most (text and image ads are not charged). Apify's free $5 credit covers it many times over.

Swap the scraper

The scrape step is a plain HTTP call to silva95gustavo/google-ads-scraper ({ "startUrls": [{ "url": ... }], "resultsLimit": N }). Any Transparency Center scraper whose rows carry the ad's YouTube link works — lexis-solutions/google-ads-scraper, solidcode/ads-transparency-scraper and others: change the actor slug and JSON body in the Scrape the Transparency Center node. The transcript step reads the run's dataset ID and deep-scans rows for the YouTube link or video ID, whatever the field names. You can also skip the scraper entirely — the transcript actor finds an advertiser's video ads itself when given advertiser names or domains.

Want the image and text ads' copy too? Pair it with steadyfetch/google-ads-creative-text-scraper — same advertiser input, extracts headlines, body copy and CTAs from the non-video creatives.

Reels swipe file — weekly hooks, AI rewrites & a Telegram digest

What it does: every Monday it loops over the creators you follow, runs steadyfetch/instagram-reel-transcript-scraper on each handle (the actor lists the newest reels itself — no scraper to chain, no login), drops every reel it has already filed with the native Remove Duplicates node, then has an AI label the hook and rewrite it for your own brand in 20 words. New rows go to Google Sheets and the five most-liked new hooks go to Telegram. A quiet week writes nothing and sends nothing.

Built for: UGC creators and social media managers keeping a swipe file they can film from.

Download the workflow JSON →

3-minute setup

  1. In n8n: Workflows → Import from File → pick instagram-reel-transcripts.workflow.json.
  2. Create an HTTP Header Auth credential (name Authorization, value Bearer YOUR_APIFY_TOKEN — token from Apify → Settings → Integrations) and select it on the HTTP Request node.
  3. Add a Groq credential on the Groq chat model node — or drop in an OpenAI / Anthropic chat model node in its place.
  4. Connect Google Sheets and Telegram.
  5. Open Config: one handle per line in creatorHandles (natgeo, @natgeo or a profile URL), plus reelsPerProfile, yourBrand (one line the AI rewrites hooks for) and telegramChatId. Optional: includeOnScreenText → true to read text off silent reels (those become charged rows).
  6. Click Test workflow, then leave the weekly Schedule Trigger on.

What you get per reel

field example
creator · reel_url · posted_at natgeo · https://www.instagram.com/reel/… · 2026-09-04T17:31Z
hook_type (AI) question / statistic / pain-point / story / bold-claim / listicle / other
content_pillar (AI) education / entertainment / behind-the-scenes / promo / story / other
emotional_trigger (AI) curiosity
cta_present (AI) true / false
rewrite_for_my_brand (AI) your hook, rewritten in ≤20 words
likes · plays · seconds · transcript_words 18,985 · 299,322 · 69.6 · 146
language · status · caption English · transcribed (only these are charged) · the post caption
charged true / false — silent reels and dead links are $0
transcript full speech-to-text, any length

Cost

steadyfetch/instagram-reel-transcript-scraper — from $7.50 / 1,000 reels on Apify's Business plan ($15.00 on the free plan), plus $0.005 per minute beyond the first 3 minutes of a long video. A 10-reel test ≈ $0.15.

Keyword opportunity scorer — a form in, scored and clustered keywords out

What it does: submit a keyword list through the built-in n8n Form Trigger (or run it manually with the demo list) → steadyfetch/keyword-search-volume-scraper returns real Google Ads Keyword Planner figures through a licensed provider — average monthly searches, competition, top-of-page bid range, 12-month trend and CPC wherever Google publishes one — then an AI labels each keyword's search intent, a topic cluster and the page type worth building. A Code node scores every keyword volume × (1 − competition/100) ÷ CPC, sorts by it, and Remove Duplicates skips anything already scored for that country in an earlier run. New rows go to Google Sheets; the ten best opportunities go to Slack.

No modelled numbers, no invented difficulty score; keywords Google has no data for come back uncharged. Switch mode to ideas to expand each keyword into new keyword ideas with the same metrics.

Built for: SEO and SEM specialists who want the list triaged before they read it.

Download the workflow JSON →

3-minute setup

  1. In n8n: Workflows → Import from File → pick keyword-search-volume.workflow.json.
  2. Create an HTTP Header Auth credential (name Authorization, value Bearer YOUR_APIFY_TOKEN — token from Apify → Settings → Integrations) and select it on the HTTP Request node.
  3. Add a Groq credential on the Groq chat model node — or drop in an OpenAI / Anthropic chat model node in its place.
  4. Connect Google Sheets and Slack.
  5. Open the On form submission node and copy its form URL — that is the plug-and-play front door (keywords, country, language). The Config node holds the demo list, mode and slackChannel for manual runs.
  6. Click Test workflow, or open the form and paste a list.

What you get per keyword

field notes
rank · opportunity_score position in the sorted list and the score behind it
intent (AI) informational / commercial / transactional / navigational
cluster (AI) a two or three word topic label
suggested_page_type (AI) blog / comparison / product / landing / tool
avg_monthly_searches Google's average; 0 means fewer than Google reports
cpc_usd average CPC, null where Google publishes none
competition · competition_index · low_top_of_page_bid_usd · high_top_of_page_bid_usd LOW / MEDIUM / HIGH, 0–100, and the bid range
trend · data_as_of rising / flat / falling and the last month Google reported
charged true / false — no-data keywords are $0

The score favours high volume, low competition and cheap clicks. Reweight it in the Score the opportunity Code node.

Cost

steadyfetch/keyword-search-volume-scraper — from $2.00 / 1,000 keywords on Apify's Business plan ($8.00 on the free plan), no minimum batch. From 16 September 2026 a run that buys fresh data also pays one $0.19 fresh-lookup fee (runs answered from the actor's 30-day cache pay none). 50 fresh keywords ≈ $0.59 on the free plan ($0.40 before that date); set the run's Maximum cost per run to at least $0.25.

Instagram competitor watch — daily posts, a viral flag & a Slack alert

What it does: every morning it loops over the competitor accounts you name, runs steadyfetch/instagram-profile-posts on each (no login, no cookies; your limit is exact, 30 means 30), drops every post it has already reported with the native Remove Duplicates node, then works out which of the new ones broke out: engagement (likes + comments) is compared with the median for that same account, so a big account and a small one are judged fairly. An AI reads each caption for theme, tone, format and CTA. Every new post is appended to Google Sheets — but Slack is pinged only when something went viral.

Built for: social media managers who want to hear about the breakouts, not about every post.

Download the workflow JSON →

3-minute setup

  1. In n8n: Workflows → Import from File → pick instagram-profile-posts.workflow.json.
  2. Create an HTTP Header Auth credential (name Authorization, value Bearer YOUR_APIFY_TOKEN — token from Apify → Settings → Integrations) and select it on the HTTP Request node.
  3. Add a Groq credential on the Groq chat model node — or drop in an OpenAI / Anthropic chat model node in its place.
  4. Connect Google Sheets and Slack.
  5. Open Config: one handle per line in competitorHandles, plus postsPerProfile, viralMultiplier (2 = twice that account's median) and slackChannel.
  6. Click Test workflow, then leave the daily Schedule Trigger on.

What you get per post

field example
account · posted_by · post_url natgeo · natgeo · https://www.instagram.com/p/DchLnq8E21N/
posted_at · type 2026-08-26T21:30:18Z · carousel
theme · tone (AI) wildlife conservation · factual
cta_present · format_guess (AI) false · carousel
summary (AI) ≤20 words
likes · comments · plays 19,990 · 140 · —
engagement · account_median_engagement · is_viral 20,130 · 8,400 · true
caption "How it started 👉 how it's going…"
charged true / false — posts we could not deliver are $0

Raise viralMultiplier to 3 for a quieter alert.

Cost

steadyfetch/instagram-profile-posts — from $0.60 / 1,000 posts on Apify's Business plan ($2.40 on the free plan). Two profiles × 30 posts ≈ $0.15 on the free plan. From 16 September 2026 each profile that delivers posts adds one $0.003 profile lookup (two profiles ≈ $0.006 more); a profile that delivers nothing pays none.

Validated

All eight workflows are gated on the latest n8n in Docker — 2.37.7 — not just parsed. Every sticky note is measured on a rendered canvas at 100% zoom (clipped text is silent: the sticky wrapper is overflow: hidden), and every workflow is executed end to end against the live Apify API from the editor's own run endpoint, with one HTTP Header Auth credential and the Google Sheets / Slack / Gmail / Telegram nodes disabled.

The five scheduled monitors are run twice: pass 1 appends the rows and sends the digest, pass 2 finds nothing new — the native Remove Duplicates node drops every row it already reported and the workflow ends on its "nothing new" branch. That is the whole point of a scheduled template, so it is proven rather than assumed.

Every workflow uses an HTTP Header Auth credential for the Apify token (no token field anywhere) and writes to a Google Sheets append node.

About

Free n8n templates for ad-creative data pipelines — Facebook ad transcripts, hooks & CTAs

Resources

Readme
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

Packages

Contributors