Archive.org
Search and access Internet Archive content, inspect analytics, Wayback captures, IIIF images, and account tasks, and manage item files, metadata, derivation, and the connected user's reviews.
Verdict
Common use cases
- Pull historical newspaper scans into research briefs
- Verify claims against archived web pages
- Download public domain audio for podcast production
- Check identifier availability before bulk uploads
- Automate metadata updates for archival collections
Integration
- Vendor
- Archive.org
- Category
- other
- Auth
- API_KEY
- Tools
- 26
- Composio slug
archive_org
Tools
- Advanced Search Archive.org Items
Search Archive.org metadata with exact Advanced Search page semantics. Use Search Archive.org Items instead when cursor-based deep paging is needed.
- Check Archive.org Upload Limit
Check whether Archive.org's ingest queue is likely to accept an upload or delete, using only the connection's LOW Authorization header.
- Check Identifier Availability
Check whether an Archive.org item identifier is available before attempting to create an item.
- Copy Item File
Copy one Archive.org file server-side to a named path in an existing destination item. This creates the destination file when absent and overwrites it when present while preserving the prior destination under history/files. The source remai
- Delete Item Filedestructive
Submit asynchronous deletion of one named file from an Archive.org item through IA-S3. An accepted response does not verify file absence; read item metadata later to verify it. This action cannot delete an item/bucket. Cascade deletion is d
- Delete My Archive.org Reviewdestructive
Permanently delete the connected Archive.org user's public review from one item. Archive.org queues the deletion and rejects the request if no such review exists.
- Derive Archive.org Item
Queue Archive.org's fixed derive.php task for an item owned or editable by the connected account. The task transforms source content into derived files asynchronously and may replace prior derivatives; this action never requests the unteste
- Download Item File
Download one public file from an Archive.org item and return a temporary downloadable file reference.
- Find Wayback Snapshots
Find the closest archived Wayback Machine snapshot for each requested URL, using one before, after, or either direction for the batch.
- Get Archive.org Task Log
Download a plain-text log for a task visible to the connected account. Optionally return only the first or last requested number of lines.
- Get Connected Archive.org User
Return the authenticated Archive.org user's identity and privileges without returning either IA-S3 credential.
- Get Detailed Archive.org Views
Get bounded geographic view analytics and, for items, top referrers over a recent window or date range. Collections support both timeframes; contributors support date ranges only and do not include referrers.
- Get IIIF Image Info
Get normalized IIIF Image API capabilities through Archive.org's fixed-host single-image convenience route or a live-proven versioned file route.
- Get IIIF Manifest
Get a compact IIIF Presentation manifest for an Archive.org item or one page of an Archive.org collection.
- Get Item Metadata
Get an Archive.org item's full record, including metadata, file inventory, public reviews, and catalog tasks, or return one selected metadata field or array slice.
- Get Item View Counts
Get all-time, last-30-day, and last-7-day view counts for one or more Archive.org item or collection identifiers from the current or legacy counting pipeline.
- Get Item View History
Get a bounded daily view history for one Archive.org item or collection, split by robot classification.
- Get My Review
Get the connected Archive.org user's review for one item, if that user has reviewed it.
- Get Task Rate Limits
Report the connected account's current task capacity and in-flight counts for one Archive.org catalog command.
- List Archive.org Tasks
Return any documented combination of the connected account's task summary, active catalog, and resource-authorized history, one page at a time.
- List Wayback Captures
Return one bounded structured CDX page of Wayback captures for a URL or host.
- Render IIIF Image
Render a bounded crop, resize, rotation, quality, and format from a fixed Archive.org IIIF image route.
- Search Archive.org Items
Search Archive.org item metadata with a Lucene-style query and return one deep-pagination page of matching items.
- Update Item Metadata
Apply JSON Patch changes to one Archive.org item owned or editable by the connected account, limited to ordinary item metadata and metadata records for named files. Remove permanently deletes the selected metadata value; move removes a valu
- Upload Item File
Create or replace one file through Archive.org IA-S3, explicitly choosing a new-item or existing-item upload. New-item uploads alone send item creation metadata. Overwrites preserve the prior file under history/files, and this action never
- Upsert My Archive.org Review
Create or overwrite the connected Archive.org user's single public review for an eligible item. The title, body, and optional rating become publicly visible on the item.
Setup
Setup guide
- 11. Log into your Archive.org account and navigate to https://archive.org/account/s3.php to generate S3-like API credentials (access key and secret key). 2. In Switchy, open your workspace settings and click 'Add Integration' under the MCP section. 3. Select 'Archive.org' from the list and paste your access key into the 'API Key' field and your secret key into the 'Secret' field. 4. Click 'Connect' — Switchy validates the credentials against Archive.org's metadata API. 5. Open any Project and type '@Archive.org search for items about Apollo 11' to confirm the connection works. 6. The MCP will return metadata records; use 'Download Item File' to pull actual content into your workspace. 7. Write operations (upload, delete, derive) require your Archive.org account to own or have edit rights on the target item identifier.
What teammates see: by default, memories from Archive.org are scoped to the Project (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Projects have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
Search Historical News
@Archive.org search for items in the 'newspapers' collection about the 1969 moon landing published between July 1969 and August 1969, return the top 10 results with download links
Download Public Domain Book
@Archive.org find the item identifier for 'The Great Gatsby' by F. Scott Fitzgerald, then download the PDF file to my workspace
Check Upload Identifier
@Archive.org check if the identifier 'company-annual-report-2024' is available for a new upload, and tell me the current upload queue status
Retrieve Archived Webpage
@Archive.org search the Wayback Machine for snapshots of 'example.com/pricing' captured in January 2020, return the closest match and its content
Derive Item Files
@Archive.org queue the derive task for item 'my-video-upload-2024' to regenerate thumbnails and MP4 derivatives, then confirm the task was accepted
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@Archive.org search for public domain recordings of Beethoven's 5th Symphony uploaded in the last 5 years, sorted by downloads
Found 47 items matching your search. Top results include 'Beethoven Symphony No. 5 - Berlin Philharmonic 2019' (identifier: beethoven-5-berlin-2019, 12,400 downloads), 'Fifth Symphony Complete Performance' (identifier: beethoven-fifth-complete-2020, 8,900 downloads), and 'Beethoven 5 Remastered' (identifier: beethoven5-remaster-2021, 7,200 downloads). Each item includes multiple audio formats (MP3, FLAC, OGG) and metadata showing license status, collection membership, and file sizes. This search used Advanced Search semantics to filter by mediatype:audio, creator field matching, and year range.
This example demonstrates metadata search across Archive.org's 40+ million items. The MCP returns structured results with identifiers you can use in subsequent file operations. Search performance depends on query complexity; broad searches may take 3-5 seconds. API key must have read permissions.
@Archive.org check if the identifier 'my-project-backup-2024' is available, then upload my local database dump to create that item with collection 'opensource'
Identifier 'my-project-backup-2024' is available. Checking upload queue capacity... Current queue status: ACCEPTING. Proceeding with upload. Created new item at archive.org/details/my-project-backup-2024 and uploaded 'database-dump.sql.gz' (245 MB). File is queued for derivative generation—MP3/PDF conversions will appear within 1-6 hours depending on queue depth. Item is now publicly accessible with metadata: title 'My Project Backup 2024', collection 'opensource', mediatype 'data'. Note: This action is permanent and cannot be undone through the API.
This example shows the full create-and-upload workflow. Archive.org uploads are asynchronous—your file appears immediately but derived formats (thumbnails, text extraction) process in background queues. The MCP checks identifier availability first to prevent collision errors. Requires API key with upload permissions and respects Archive.org's rate limits (typically 100 requests/minute).
@Archive.org download the highest-quality audio file from item 'grateful-dead-1977-05-08' and summarize its technical metadata—sample rate, bitrate, duration
Downloaded 'gd1977-05-08-sbd-miller.flac' (892 MB) from the Grateful Dead concert archive. Technical analysis: FLAC format, 24-bit/96kHz sample rate, average bitrate 4,608 kbps, duration 2:18:47 (138 minutes, 47 seconds). File metadata shows this is a soundboard recording from Cornell '77, digitized from master tapes. The item contains 8 additional formats including MP3 V0, Ogg Vorbis, and a 256kbps MP3. This FLAC represents the archival-quality preservation master. File is now available as a temporary reference for further processing.
This example pairs Archive.org's file download with AI analysis of technical properties. The MCP retrieves public files without authentication, but API-key connections get higher rate limits. Large files (>500 MB) may take 30-60 seconds to transfer. The temporary file reference expires after your session, so extract needed data immediately.
Use-case deep-dives
When Archive.org MCP makes sense for long-term dataset preservation
A 3-person research team publishing open datasets from a university lab needs permanent public hosting with zero ongoing cost. Archive.org MCP is the right call here because it handles the full lifecycle: check identifier availability before upload, verify ingest queue capacity, push files via IA-S3, and trigger derivative generation for multi-format access. The 26 tools cover edge cases like server-side file copying when you need to reorganize an archive post-publication. The threshold: if your team needs private staging or embargoed releases, Archive.org's public-first model won't work—you'd need S3 or institutional repositories with access control. For truly public,永久 archival with metadata search built in, this MCP eliminates the plumbing between your workflow and the Internet Archive.
Archive.org MCP for one-time bulk media uploads
A solo podcaster moving 200 episodes off a dying host to Archive.org hits the sweet spot for this MCP. The workflow is linear: check upload limits, batch-create items with metadata, push audio files, queue derivative generation for streaming formats, then verify completion by reading item metadata. The MCP's delete and copy tools handle mistakes without re-uploading gigabytes. Where it breaks down: if you're publishing new episodes weekly and need automated RSS feed updates, Archive.org's async derive queue (sometimes 24+ hours) makes it a poor fit for production pipelines. This is a migration tool, not a live publishing stack. For back-catalog preservation where you control the timeline, the MCP turns a multi-day manual slog into a scripted afternoon.
When Archive.org MCP supports collaborative digital collections
A 6-person volunteer team curating a local history collection on Archive.org uses this MCP to triage uploads and moderate reviews. The advanced search tool lets them filter items by uploader, date, or metadata fields to spot duplicates or policy violations. Delete item file and delete review tools handle takedown requests without touching the Archive.org web UI. The copy file tool reorganizes misplaced scans across items. The limit: Archive.org's API key auth means each team member needs their own credentials—there's no shared service account, so audit trails live in Archive.org's logs, not your workspace. If your moderation SLA is under 2 hours, the async derive queue will frustrate you. For weekly triage cycles and teams comfortable with individual API keys, this MCP beats manual clicking through hundreds of items.
Frequently asked
What can the Archive.org MCP do in Switchy?
It lets your team search Archive.org's catalog, upload files to new or existing items, check identifier availability, and manage metadata—all without leaving your AI workspace. You can also queue derive tasks to generate formats like PDFs or MP3s from source files, download public files, and delete reviews or files you own. It's built for teams that archive content programmatically or need to pull historical data into their workflows.
Do I need special permissions to connect Archive.org?
You need an Archive.org account with API key credentials. The MCP checks your upload quota using the LOW authorization level, so you don't need admin rights to the entire archive—just valid credentials for your own account. If you're uploading to items you don't own, those items must allow edits from your account, which typically means they're community-editable or you're listed as a contributor.
Can it upload files larger than 100 MB?
Yes, but Archive.org enforces its own ingest queue limits based on your account's standing and current load. The MCP includes a tool to check whether the queue will likely accept your upload before you try. If the queue is full or your account is rate-limited, you'll get a rejection—this isn't a Switchy restriction, it's Archive.org's infrastructure protecting itself from overload.
How does this compare to uploading through Archive.org's website?
The MCP uses Archive.org's IA-S3 API, which is faster for bulk operations and lets you script uploads, metadata edits, and file copies without clicking through web forms. You lose the web UI's visual feedback and preview features, but you gain the ability to chain actions—like checking identifier availability, uploading, then queueing a derive task—inside a single AI conversation or automation.
Who on the team should connect this MCP?
Whoever owns the Archive.org account you want to use. The API key ties actions to that user's upload quota and item ownership, so if your team shares an organizational Archive.org account, connect it once and share the Switchy workspace. If individuals upload under their own names, each person connects their own key. Either way, uploads and edits appear in Archive.org attributed to the connected account.