Index of every page on this site: llms.txt
All skills
exa-labs avatar

/build-with-exa

@9d220cc
by Exa Labsexa-labs/agent-skills75 stars
8
Run it yourself

Build applications and agents with Exa's API: search, contents extraction, answer, Agent API, monitors, websets, OpenAI-compatible endpoints, and exa-py/exa-js SDKs. Use when choosing Exa endpoints, writing Exa API calls, integrating semantic web search or research into products, or debugging Exa request shapes.

Use this Skill: https://skilld.dev/gh/exa-labs/agent-skills/build-with-exa

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

referencessnapshot.md

≈1.3k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Exa Snapshot Reference

Historical page versions via snapshotAsOf on POST /contents and POST /search.

Canonical Docs Links

  • Base docs URL: https://exa.ai/docs
  • Exa Snapshot: /search/snapshot
  • Contents reference: /reference/get-contents
  • Search reference: /reference/search

Contents

  • Overview
  • Contents at a datetime
  • Search at a datetime
  • How snapshots work
  • Snapshot vs freshness
  • Limits
  • Critical pitfalls

Overview

Exa Snapshot keeps stored versions of the pages Exa has crawled. Sending snapshotAsOf (an ISO 8601 date-time) pins a request to that instant: Exa returns the newest stored version of each page at or before it and never fetches the live page.

Use Snapshot when the task needs page content as it was, not as it is:

  • backtesting agents against what the web said at the time
  • reproducible evals whose inputs must not drift between runs
  • comparing earlier versions of docs, pricing pages, policies, or filings

"Snapshot" in Exa's product vocabulary always means this historical retrieval. A plain /contents call returns the current page; it is not a snapshot.

Contents at a Datetime

snapshotAsOf is a top-level field on /contents, next to urls and the content fields.

POST https://api.exa.ai/contents
{
  "urls": ["https://en.wikipedia.org/wiki/2026"],
  "snapshotAsOf": "2026-06-01T00:00:00Z",
  "text": true
}
from exa_py import Exa

exa = Exa(api_key="YOUR_EXA_API_KEY")
result = exa.get_contents(
    ["https://en.wikipedia.org/wiki/2026"],
    snapshot_as_of="2026-06-01T00:00:00Z",
    text=True,
)
import Exa from "exa-js";

const exa = new Exa();
const result = await exa.getContents(["https://en.wikipedia.org/wiki/2026"], {
  snapshotAsOf: "2026-06-01T00:00:00Z",
  text: true,
});

URLs with no stored version at or before the cutoff are omitted from results and reported in statuses with "status": "error" and "tag": "CONTENT_NOT_CACHED". Served versions carry "source": "cached".

Search at a Datetime

On /search, snapshotAsOf goes inside contents, like every other content option.

POST https://api.exa.ai/search
{
  "query": "latest stable Python release notes",
  "numResults": 3,
  "contents": {
    "snapshotAsOf": "2026-07-01T00:00:00Z",
    "highlights": true
  }
}
result = exa.search(
    "latest stable Python release notes",
    num_results=3,
    contents={"snapshot_as_of": "2026-07-01T00:00:00Z", "highlights": True},
)

Exa discovers candidate URLs with its current retrieval signals, then keeps only pages with a stored version at or before snapshotAsOf and serves that version. The cutoff bounds the content, not the ranking: treat the results as evidence bounded by the datetime, not as a reconstruction of what a search would have ranked then.

How Snapshots Work

Field Where Meaning
snapshotAsOf /contents top level; /search inside contents ISO 8601 date-time cutoff. Exa returns the newest stored version at or before this instant.

For both endpoints:

  • Returned page content comes from that stored version.
  • Title, author, publication date, text, highlights, and summaries are generated only from that version.
  • Pages without an eligible version in the retention window are omitted.

Snapshot vs Freshness

snapshotAsOf and maxAgeHours answer opposite questions and are never sent together.

Need Field
The page as it was at a past instant snapshotAsOf
The page as it is now, at most N hours old maxAgeHours (see references/contents.md)
Results published within a date range startPublishedDate / endPublishedDate (see references/search.md)

Neither maxAgeHours: 0 (force a live crawl) nor a published-date filter returns an earlier version of a page.

Limits

  • Pay as you go can query a rolling 5-month window. Please reach out to our team for access to our entire historical coverage.
  • Available on pay as you go at 10 QPS. After 100 Snapshot requests, contact sales (https://exa.ai/contact/sales) to continue.

Critical Pitfalls

  1. Do not answer "snapshot" with a plain /contents request or with maxAgeHours; only snapshotAsOf returns a stored past version.
  2. On /contents the field is top level; on /search it is nested inside contents.
  3. Do not combine snapshotAsOf with maxAgeHours, livecrawl, livecrawlTimeout, or subpages; the request is rejected with INVALID_REQUEST. On /search, use auto or lite. deep-lite requires account access. Exa Snapshot supports all Search categories through the category parameter. deep and deep-reasoning are not supported.
  4. Search results under snapshotAsOf are content-bounded, not ranking-bounded; say so when the task is "what would a search have returned then".
  5. Expect omissions: pages without a stored version before the cutoff are dropped (CONTENT_NOT_CACHED on /contents), so check statuses and result counts.

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

1 warning2d3 checks · Risk SAFE
  • Gen Agent Trust Hub3d

    The skill provides official documentation and SDK guidance for the Exa API. It includes instructions for installing and using official Python and TypeScript libraries and provides best practices for semantic search and AI research workflows.

  • Socket2d

    No alerts

  • Snyk3d

    Risk: MEDIUM · 1 issue

Signed by skilld at 9d220cc. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 4 days ago.

Activeupdated 5 days ago
metadata
{
  "author": "Exa",
  "version": "0.3.0",
  "docs": "https://exa.ai/docs"
}

README badge

README badge for exa-labs/agent-skills/build-with-exa