Integration

Unstructured Transform integration

Unstructured Transform turns documents (PDF, Word, PowerPoint, and images such as PNG, JPEG, TIFF, BMP or HEIC, up to 100 MB) into structured data an agent can use. Parse a document into Markdown or into typed elements (titles, narrative text, tables, with page numbers and optional coordinates and table HTML), optionally extracting fields described by a JSON Schema in the same job; extract fields later from a parse that already exists; upload a file once to reuse it by id for 24 hours; and list, read, cancel or delete jobs. Live progress streaming (server-sent events) is not exposed: a long parse or extraction that outlives the wait comes back as a queued job to read with get_job. Authentication is an API key created at transform.unstructured.io (API keys page), sent in the unstructured-api-key header. Unstructured Business accounts (platform.unstructured.io) use a different API URL and key, not covered here.

Connect Unstructured Transform How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in Unstructured Transform.Built from the 9 Unstructured Transform tools below, on the connection you add once.

What you can do with Unstructured Transform

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Job, 4 operations

cancel_jobCancel jobPOSTRequests cancellation of a queued or processing job and returns the job as it stands (cancellation_requested true); read it again with get_job until its status is cancelled.
delete_jobDelete jobDELETERemoves a finished job (completed, failed or cancelled) from the job listings. Deletion is asynchronous and idempotent; a job still running cannot be deleted.
get_jobGet jobGETReturns a parse or extraction job's status and, once available, its content in result (markdown or elements, and extracted_data for extraction fields). Use it to read a job that parse_document or extract_fields returned still queued.
list_jobsList jobsGETLists the parse and extraction jobs visible to this API key, newest first, with their status, operation, file and whether a result can still be retrieved. Jobs whose result has expired are omitted. Page with cursor and next_cursor.

File, 3 operations

delete_uploaded_fileDelete uploaded fileDELETEDeletes an uploaded file from scratch storage before it expires. Idempotent: an id that is already deleted or was never valid also succeeds.
download_uploaded_fileDownload uploaded fileGETDownloads the original bytes of a file uploaded earlier with upload_file, as a stored file, while it has not expired.
upload_fileUpload filePOSTUploads a document to scratch storage for 24 hours and returns its file_id, detected MIME type and expiry time, so the same file can be parsed or extracted several times without sending it again.

Document, 1 operation

parse_documentParse documentPOSTParses one document (PDF, DOCX, PPTX or an image, up to 100 MB) sent as a file or by an uploaded file_id, and returns Markdown (default) or a list of typed elements. With schema it also extracts those fields into extracted_data in the same job.

Field, 1 operation

extract_fieldsExtract fieldsPOSTExtracts structured fields, described by a JSON Schema, from a document that parse_document has already parsed (identified by parse_id), without parsing it again. The values arrive in extracted_data[].data, with per-field citations in field_metadata.

Use Unstructured Transform from Claude Code, Cursor or Codex

Connect Unstructured Transform in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its Unstructured Transform operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect Unstructured Transform

Create your workspace
Sign up free. Nothing to install.
Connect Unstructured Transform once
Add your Unstructured Transform connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.