Exa Snapshot

Exa Snapshot keeps stored versions of pages Exa has crawled. Send snapshotAsOf to pin the request to a datetime. Use this to backtest agents, run reproducible evals, and compare earlier versions of docs, pricing pages, policies, and filings.

Exa Snapshot is available on pay as you go at 10 QPS, with a rolling 5-month index window. After 100 requests, talk to sales to continue.

Pin contents to a datetime

Add snapshotAsOf at the top level of a /contents request.

from exa_py import Exa

exa = Exa()

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

print(result.results[0].text[:300])

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
  }
);

console.log(result.results[0].text.slice(0, 300));

curl -s -X POST "https://api.exa.ai/contents" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $EXA_API_KEY" \
  -d '{
    "ids": ["https://en.wikipedia.org/wiki/2026"],
    "snapshotAsOf": "2026-06-01T00:00:00Z",
    "text": true
  }'

Exa returns the newest stored version at or before that datetime.

Example response

{
  "requestId": "c05151f7df9cd9d8785e0acf0935355d",
  "results": [
    {
      "id": "https://en.wikipedia.org/wiki/2026",
      "url": "https://en.wikipedia.org/wiki/2026",
      "title": "2026",
      "author": null,
      "text": "2026\n\n2026 (MMXXVI) is the current year, and is a common year starting on Thursday of the Gregorian calendar...",
      "image": "https://upload.wikimedia.org/wikipedia/commons/thumb/9/93/..."
    }
  ],
  "statuses": [
    {
      "id": "https://en.wikipedia.org/wiki/2026",
      "status": "success",
      "source": "cached"
    }
  ]
}

IDs with no eligible version are omitted from results and reported in statuses with "status": "error" and "tag": "CONTENT_NOT_CACHED".

Search at a datetime

On /search, put snapshotAsOf inside contents.

from exa_py import Exa

exa = Exa()

result = exa.search(
    "latest stable Python release notes",
    num_results=3,
    contents={
        "snapshot_as_of": "2026-07-01T00:00:00Z",
        "highlights": True,
    },
)

for r in result.results:
    print(r.title, r.url)

import Exa from "exa-js";

const exa = new Exa();

const result = await exa.search("latest stable Python release notes", {
  numResults: 3,
  contents: {
    snapshotAsOf: "2026-07-01T00:00:00Z",
    highlights: true
  }
});

for (const r of result.results) {
  console.log(r.title, r.url);
}

curl -s -X POST "https://api.exa.ai/search" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $EXA_API_KEY" \
  -d '{
    "query": "latest stable Python release notes",
    "numResults": 3,
    "contents": {
      "snapshotAsOf": "2026-07-01T00:00:00Z",
      "highlights": true
    }
  }'

Exa discovers candidate URLs, then keeps only pages with a stored version at or before snapshotAsOf.

Example response

{
  "requestId": "211fc1f57b87a792de082309ef3bce95",
  "results": [
    {
      "id": "https://docs.python.org/3/whatsnew/changelog.html",
      "url": "https://docs.python.org/3/whatsnew/changelog.html",
      "title": "Changelog — Python 3.14.6 documentation",
      "highlights": [
        "Changelog — Python 3.14.6 documentation\n...\n## Python 3.14.6 final¶\n...\nRelease date: 2026-06-10"
      ],
      "image": "https://docs.python.org/3.14/_images/social_previews/..."
    },
    {
      "id": "https://docs.python.org/3/whatsnew/index.html",
      "url": "https://docs.python.org/3/whatsnew/index.html",
      "title": "What's New in Python — Python 3.14.6 documentation",
      "highlights": ["What's new in Python\n...\n- Python 3.14.6 final\n- Python 3.14.5 final"]
    },
    {
      "id": "https://docs.python.org/3/whatsnew/3.14.html",
      "url": "https://docs.python.org/3/whatsnew/3.14.html",
      "title": "What's new in Python 3.14 — Python 3.14.6 documentation",
      "highlights": ["Python 3.14 is the latest stable release of the Python programming language..."]
    }
  ]
}

How snapshots work

Field	Where	Meaning
`snapshotAsOf`	Request	Datetime 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 5-month window are omitted.

On Search, the cutoff bounds content, not ranking. Exa still uses current retrieval signals to discover candidate URLs. Use the results as evidence bounded by snapshotAsOf, not as an exact reconstruction of what a search would have ranked at that time.

Limits and compatibility

Access, rate limit, and lookback

Pay as you go includes 10 QPS and a rolling 5 months of index access. A snapshotAsOf older than that window is rejected. After 100 requests, talk to sales to continue.

Historical requests use stored content

Do not combine snapshotAsOf with options that can reach the live web or expand to other pages. Omit livecrawl, livecrawlTimeout, maxAgeHours, and subpages entirely; requests that set any of them alongside snapshotAsOf are rejected with INVALID_REQUEST.

Supported Search requests

Exa Snapshot on Search supports auto, fast, and instant. It does not support deep-lite, deep, or deep-reasoning.Exa Snapshot does not support the category parameter on Search.

Common uses

Use Exa Snapshot when the task depends on what Exa had stored by a specific datetime:

Backtest an agent without exposing it to later page updates.
Run an evaluation against a repeatable content boundary.
Compare earlier versions of documentation, pricing, policies, or filings.

Get Started

Search

Agent

Other APIs

SDKs

Integrations

Admin

Websets

Exa Snapshot

Pin contents to a datetime

Search at a datetime

How snapshots work

Limits and compatibility

Common uses

​Pin contents to a datetime

​Search at a datetime

​How snapshots work

​Limits and compatibility

​Common uses

Pin contents to a datetime

Search at a datetime

How snapshots work

Limits and compatibility

Common uses