Find the post where a writer covered a topic, list everything they've published and read the free ones in full without scrolling an archive page. With the Firecrawl plugin, ChatGPT can search a Substack archive through Firecrawl Alexandria, using the archive and search that Substack serves for each publication.
Best for
- Readers hunting for the post where a newsletter writer first made an argument.
- Researchers pulling every post on one subject from a writer's back catalog.
- Writers and editors checking what a newsletter has already covered before pitching or publishing.
Plugins and data providers
Plugin: the Firecrawl plugin. Lets ChatGPT reach Substack archives in Firecrawl Alexandria, with each lookup billed in Firecrawl credits.
Data providers in Firecrawl Alexandria
| Provider | Tool | Use it for |
|---|---|---|
| Substack | Archive | One page of up to 50 posts, sorted newest, most popular or pinned, with title, subtitle, link, publish time, free or paid, word count, reactions, comments, restacks and byline. Page further back with an offset. |
| Substack | Search | Runs Substack's own archive search for one publication and returns the same post entries as the archive. |
| Substack | Post | One post by link, with its body text. Free posts come back complete; paid posts return the public preview, marked as paywalled with the share of words visible. |
Explore Firecrawl Alexandria data providers across categories
Starter prompt
Put the newsletter and the subject you care about where Noahpinion and tariffs are:
@firecrawl Search the Noahpinion Substack archive for posts about tariffs
using Firecrawl Alexandria. Tell me what each tool costs before running it.
Give me:
- Every matching post with its title, date, link and whether it's free
or paid.
- The three most discussed, by comments and reactions.
- A short summary of the newest free post, read from its full text.
Mark any paid post as a preview, and only quote text the tool returned.Before you start
- Get the Firecrawl plugin. The install button under these steps opens its page inside ChatGPT.
- Link a Firecrawl account. Credits for each Substack lookup come out of it, and a new one takes a minute to sign up for.
- Know the newsletter's address. Use the slug from a substack.com address (garymarcus for garymarcus.substack.com). Of the newsletters on their own domains, only Astral Codex Ten, Lenny's Newsletter and Noahpinion work today.
Limitations of using ChatGPT to search a Substack archive
A Substack can hold years of posts, some free, some locked and some never indexed. Asking ChatGPT to search one runs into five problems.
- Search returns a sample of posts. OpenAI warns that what ChatGPT's search returns, citations included, may be missing pieces, out of date or wrong, and that no site is promised a spot in it (OpenAI). A few posts on a topic don't prove the writer never covered it elsewhere.
- Old posts can go behind a paywall. Substack lets writers auto-paywall their back catalog, so free subscribers and new readers hit a paywall on older free posts, and posts the writer paywalled by hand stay that way (Substack). Without a paid subscription, those posts can't be read in full.
- Private newsletters stay out of search engines. For a private publication, Substack says the Welcome page may be indexed but the posts won't be (Substack). Search can find the newsletter's name and none of what it published.
- Some pages are out of reach. ChatGPT may not get information from a site because of technical issues, paywalls or robots.txt settings, and without tools the model only knows what it saw before its training data ends (OpenAI). This week's post won't be in the model, and last year's may not be in the search results.
- Invented posts and quotes. The same OpenAI page lists fabricated quotes, citations and references to sources that don't exist among ChatGPT's mistakes. A believable title from a writer you follow is easy to make up, so open the link before you repeat a quote.
How it works
Four pieces do the work when you ask about a newsletter:
- Firecrawl Alexandria: a set of structured data tools that agents can call by name. Over 100 official providers are licensed into it, and they're compensated whenever their data is pulled. Firecrawl's Web, Research, Developer and Government indexes sit beside them, along with public sources such as Substack. Across 845 benchmark tasks, holding model and prompt constant, answers from agents with Firecrawl Alexandria rated 21% better than answers from agents with only the default web tools (Firecrawl's evals).
- The publication's own archive: the Substack tools read the archive and search that the newsletter itself serves, so results cover posts web search may never surface. ChatGPT pages through older posts 50 at a time; in our Oct 9, 2026 API test, Noahpinion's archive reached back to posts from 2022.
- What's free and what isn't: every post entry says whether it's free or paid. Reading a paid post returns only the preview a logged-out visitor sees, flagged as paywalled, with the visible share of words.
- Read-only, priced up front: ChatGPT shows the credit cost before a call, and the tools only read public pages. Nothing is posted, liked or subscribed to.
Searching a Substack archive in ChatGPT with and without Firecrawl Alexandria
| No plugin | Through Firecrawl | |
|---|---|---|
| Which posts you see | Whatever the search engine ranks, often a few | The publication's archive, paged back through older posts |
| Finding a topic | Depends on how the posts were indexed | Substack's own search for that publication |
| Free or paid | Often unclear until you click | Marked on every post, with a paywall flag |
| Reading a post | A page fetch that may fail or stop at the paywall | Full text for free posts; the public preview for paid ones, with the visible share stated |
| Popularity | Rarely shown in search results | Reactions, comments and restacks per post, and a most-popular sort |
| Source | Hard to tell what was missed | Substack's public archive, one of the sources in Firecrawl Alexandria, whose catalog lists 100+ providers covering topics from media to public records. Explore the catalog |
| New posts | You check back yourself | ChatGPT can set up a Firecrawl Monitor on the newsletter's archive page and alert you by email or Slack when it changes |
Build your own Substack archive workflow
Shape the prompt around what you're after. As a researcher tracing an argument:
@firecrawl Search the garymarcus Substack for posts about scaling. List each
match with date, title and link, then read the two earliest free posts in
full and summarize how the argument started.As a writer sizing up a newsletter before pitching it:
@firecrawl Show the 20 most popular posts from Lenny's Newsletter with
reactions, comments, word count and whether each is free or paid. Which
topics come up most often?Name the publication, say whether you want search results, the newest posts or the most popular ones, and tell ChatGPT which posts to open in full.
Take it further
Catch up on a newsletter
@firecrawl List the 10 newest posts from Astral Codex Ten with dates and
whether each is free or paid, then summarize the newest free one.Go deep into the back catalog
@firecrawl Page through the Noahpinion archive and list every post from
2023 with its title and link.Check what a paywall hides
@firecrawl Open https://www.noahpinion.blog/p/the-second-trump-presidency-is-a
and tell me how much of the post is visible without a subscription.Get told about new posts
@firecrawl Monitor https://www.noahpinion.blog/archive every day and email me
when a new post appears.See the tools and prices
@firecrawl What tools can search or read a Substack newsletter? Show the
price of each one before running anything.What it costs
Browsing Firecrawl Alexandria and reading prices is free. Credits are deducted when a Substack tool actually runs, and Firecrawl pricing lists each plan's allowance.
| Substack tool | What it returns | Credits per call |
|---|---|---|
| Archive | One page of up to 50 posts, newest, top or pinned first | 5 |
| Search | Posts matching a query in one publication | 5 |
| Post | One post's text: complete if free, the preview if paid | 5 |
Paging back through 300 posts takes six archive calls, or 30 credits. After each answer, ChatGPT names the tool it called and what that cost.
[ Substack archives in ChatGPT ]
Install the Firecrawl plugin and ask ChatGPT to search any Substack's back catalog. Finding the tool is free.
Related use cases
- How to search podcast transcripts in ChatGPT: many newsletter writers also host shows, and this finds what they said on air.
- How to get real-time stock news in ChatGPT: check the market headlines behind a finance newsletter's latest take.
- Nasdaq.com: full text and exact publish times for stories in its news feeds, useful for checking the market news a post reacts to.
Coverage limits
- Paid posts are previews. Paid text stays at what an anonymous reader sees. In our Oct 9, 2026 API test, a paid Noahpinion post showed 598 of its 2,933 words.
- Supported addresses. Publications on substack.com addresses work by slug. Custom domains are limited to Astral Codex Ten, Lenny's Newsletter and Noahpinion for now.
- No date filter. The tools sort by newest, most popular or pinned and page by offset, so finding posts from a given year means paging back to it.