Resources/ Prompt libraryTry Autumn ↗Every prompt here was run in Autumn and kept because its rows held up when we checked the sources. Change the inputs, pick your columns, and copy it into a task or your agent.
For one person or company you already know, a type word and a name does the job. Autumn picks the fields itself; in our tests, long briefs that listed every field came back no richer.
You know who it is and want everything public about them in one record.
Looks inLinkedInPersonal sitesTalks and podcastsPressAdd the company after the name when the name is common. A long brief listing the fields came back no richer than this one line.
Before a call, a deal or a partnership: one record with the company's people, money, product and news.
Looks inCompany siteFunding announcementsPressJob postsPut the domain in brackets so it researches the right company. Our longer brief asking for each field by name came back with fewer fields than this.
A signup or a target account where all you have is the domain.
Looks inCompany siteFunding announcementsPressJob postsA bare domain is enough. It resolves the company first, then builds the same full record as a company profile.
You sell to teams that run a given tool, or you need to know what a company already pays for.
Looks inDNS recordsSubprocessor listsJob postsSite codeIt reads DNS records, subdomains, subprocessor lists, job posts and the site's own code, so tools show up with proof, not a vendor's guess. Versions appear only where the site's code exposes them.
| tech_name | tech_category | evidence |
|---|---|---|
| Stripe | payments | Listed on Ramp's subprocessor page |
| HubSpot | crm | Email sending record in DNS |
| Google Workspace | productivity | Email sending record in DNS |
| HackerOne | security | Domain verification record |
| Linear | productivity | Domain verification record |
| Next.js 16, React 19 | frontend | Ramp's own JavaScript bundles |
Part of what came back for ramp.com. Every tool comes with the evidence that shows the company uses it.
You need the real buying path or a team's structure, not a flat list of titles.
Looks inLinkedInTeam pagesPressAsk for the org chart, not for "reports to" lines in job posts. Most posts never name a manager; the org chart prompt builds the tree from profiles, team pages and press, with evidence for each line.
Each reporting line comes with the source it was read from.
When the answer is a list, or a match you will act on, the rules do the work. Open a prompt, swap in your inputs, drop the columns you don't need and copy it.
TaskFind 15 companies incorporated in Delaware between 2026-10-16 and 2026-10-30 that look like software or AI startups.
RulesOnly list a founder when a public source ties that person to this specific company. Leave a cell blank rather than guess.
ColumnsColumns: entity_name, incorporation_date, domain, founders, source_urls.
For one person or company you already know, one line beats a long brief.
Profile a person →Ask for the thing you want, not a proxy for it: the org chart, not reporting lines in job posts.
Map who reports to whom →Make it state the criteria before it searches, or every row gets the same reason.
Find lookalikes →Name what counts as proof, and ask it to label how it knows.
Find work emails →Let it say unclear. A blank you can see beats a guess you can't.
Students building in public →Point it at the primary source: the company's own careers page, the state registry.
Companies hiring for a role →The whole library is published as JSON and Markdown, so an agent can fetch the current version of a prompt by name instead of someone rewriting it.
LIBRARY=https://www.autumn.ai/resources/prompt-library
SLUG=new-delaware-companies
# 1. Fetch the current prompt by its slug
PROMPT=$(curl -s $LIBRARY/prompts.json \
| jq -r --arg s "$SLUG" '.prompts[] | select(.slug == $s) | .prompt')
# 2. Start an Autumn task with it
curl -sS https://api.autumn.ai/task \
-H "X-API-Key: $AUTUMN_API_KEY" \
-H "Content-Type: application/json" \
-d "$(jq -n --arg p "$PROMPT" '{prompt: $p}')"
We ran each prompt in Autumn in October 2026 and read the rows against their sources. Prompts that came back padded or generic were rewritten and run again; the lessons above come from those reruns. Example rows show public companies and executives only.
Autumn reads the public web, from registries to job posts, and returns a table with a source on every cell.
Try Autumn ↗