Libraries are the second-cheapest programmatic format after directories. One template plus a spreadsheet of terms, emails or prompts yields hundreds of pages. Ten of the competitors we crawled run at least one library of 100+ pages. Content quality is uneven. The same template that makes 600 pages can also copy one bug onto all 600.
The format is also changing. In 2024 and 2025 these libraries were email templates and glossaries. In 2026 they are prompt libraries, Claude skills and automation recipes: lemlist's /claude-skills/ (38 pages), Amplemarket's /skills/ (37), La Growth Machine's /claude-skills/ (14), and n8n/Make/Clay template galleries at HeyReach and Salesforge. Buyers no longer search only for words to send. They also search for things their agent or workflow can run.
The libraries, counted
Unique English pages from sitemaps on 5 October 2026. Translated copies are in brackets.
| Vendor | Library | Pages | Sample and observations |
|---|---|---|---|
| Instantly | /prompt/ | 329 | Every page carries a December 2025 lastmod (one batch). Pages declare Instantly's own 4.8/3,838 AggregateRating. Some slugs are run-on prompts ("1-identify-digital-transformation-needs-for-sales-teams-2-generate-…"), some are hash-suffixed duplicates ("…-first-line-generator-sales-f2a89") |
| lemlist | /email-templates/ | 239 (956 with fr/de/es) | Sequences credited to "outbound experts". Both sampled meta descriptions say "The open rate of this campaign is undefined and the reply rate undefined" |
| lemlist | /prompt-templates/ | 149 (596) | All three sampled pages share one title, "Cold Outreach Prompt Templates", and one meta description |
| lemlist | /claude-skills/ | 38 (152) | "Website Scraper", "Trigger Finder", "Copywriting Analyzer": agent skills as landing pages |
| Hunter | /templates/ | 574 | Broad, beyond sales: recruiting, SEO link-building, interview invitations |
| Hunter | /prompts/ | 114 | Prompt hub by use case |
| PlusVibe | /glossary/ | 641 (each listed twice in its sitemaps) | Generic business terms: "amortization", "accounts payable", "agile methodology". All stamped May 2025 |
| Clay | /glossary/, /templates/ | 268, 101 (536, 198 with fr) | GTM terms; templates open a Clay table |
| Amplemarket | /glossary/ | 283 | Includes seven voice-AI entries ("voice cloning", "expressivity", "text-to-speech") unrelated to sales |
| Salesforge | /glossary/, /cold-email-items/, /outbound-strategy/ | 147, 332, 843 (586 for glossary with de/fr/pt) | "Outbound strategy for |
| Woodpecker | /glossary/, /templates/ | 104, 94 | Hand-written glossary; templates by niche (real-estate follow-ups) |
| La Growth Machine | /glossary/ | 165 (×6 languages) | Sales and LinkedIn terms |
| Smartlead | /prompts/, /email-templates/ | 75, 56 | Campaign prompts ("3-step SaaS outreach campaign") |
| Apollo.io | /what-is/ | 124 | Job-role definitions ("scrum master", "accountant"): the persona side of its database |
Translations inflate these counts. lemlist's 956 email templates are 239 templates in four languages, and its 596 prompt templates are 149. PlusVibe's "1,282" glossary is 641 terms listed twice. Instantly's 329 prompts are English only.
Quality: what the templates reveal
We fetched samples from each library (5 October 2026).
- PlusVibe's glossary reads as unedited LLM output. What is Amortization? runs about 1,800 words that open "Amortization is a fundamental financial concept that plays a crucial role…" and promise "In this article, we will delve into the details". Amortization has nothing to do with cold email. A 641-page glossary of off-topic terms is close to the "many pages … without adding value" example in Google's scaled content abuse policy.
- lemlist's template data never reached the page. The email-template meta description is built from fields that come out "undefined" (example). The prompt templates share one title tag, so to a search engine 149 pages announce themselves identically. The cause looks like one data pipeline that failed, not 300 bad pages, but search engines see the failure on every page.
- Woodpecker's glossary is the cleanest. Feedback loop has a unique title, an explicit robots directive and about 1,000 words. It is small (104 terms) and on topic.
- Amplemarket's glossary drifted off topic. A sales platform defining "voice conversion" suggests a term list pulled from an AI-voice product.
Do libraries rank?
Three checks with our search tool (indicative, not Google):
| Query | Result | Read |
|---|---|---|
| "cold email template to invite podcast guests" | lemlist #1 (a landing page) and two lemlist /email-templates/ pages in the top six; Reply.io, Smartlead's blog. A test.reply.io copy of Reply's template page also appears | Template libraries win long-tail "template for X" queries |
| "what is an email feedback loop" | Zoho ZeptoMail, Drip, LeadMagic, Mailgun, Mailrelay, SendGrid docs, KumoMTA docs | Deliverability definitions go to ESPs and MTAs, not sequencers. Woodpecker's page did not appear |
| "ChatGPT prompt personalized cold email first line generator" | Docket, a prompt pack, LeadHaste, Techpresso, Tella. No Instantly /prompt/ page | Instantly's 329 prompt pages did not surface for the query they are written for |
The pattern: templates rank, glossaries mostly do not, prompts are open. Template queries carry intent ("I need words for this situation"), and a library with hundreds of situations covers the long tail. Glossary queries go to sites with topical authority, which for deliverability terms means email infrastructure companies. Prompt queries are new enough that small specialist sites still lead.
No vendor reports traffic or sign-ups from these libraries. Semrush's free pages do not break traffic out by folder. We could not test whether AI assistants cite these pages; see Who AI answer engines recommend.
Why vendors keep building them
- Internal links and freshness. Every template page links to sign-up, and bulk re-stamps of
lastmodkeep the sitemap looking fresh: Instantly's 329 prompts in December 2025, PlusVibe's glossary in May 2025. - Partner content. lemlist's templates and Clay's templates credit outside experts, who share them. The library turns into a distribution channel for the partner programme (Distribution).
- Product as content. A Claude skill or n8n template is both a page and an importable asset. Copying it into a workflow is product adoption, not just a visit.
What this means for an entrant
- Skip the generic glossary. Hundreds of definitions of "A/B testing" exist, ESPs own deliverability definitions, and an off-topic glossary is a policy risk. Write 30–50 terms where you have something to add: EU legal terms for cold email (Abmahnung, legitimate interest, soft opt-in, Impressum) in English and German, linked to the EU/EEA country matrix.
- Build a reply-handling library, not an email library. "How to answer 'not now, try in Q3'", "How to handle an Auskunftsersuchen (GDPR access request) from a prospect", "Out-of-office replies: when to resume". About 150 situations, each with the reply template, the classification your The reply desk assigns and the follow-up timing. No competitor covers the inbox side.
- Ship the library as importable assets. Each reply playbook should also exist as an n8n/Make template and as a Claude skill or MCP prompt. That matches where libraries are heading in 2026 and puts your product inside the buyer's agent (llms.txt, pricing.md and AI crawlers).
- Validate templates before publishing. Check unique titles, no "undefined" fields and no off-topic terms. lemlist and Amplemarket show that one pipeline bug repeats on every page.
- Localise only the legal and reply libraries. Product templates can stay English. The German reply-handling and legal pages are the ones local buyers search for in their own language (The keyword landscape).