Async Document Lifecycle

Once a file has been submitted via POST /files/v1/uri, POST /files/v1/jobs, or POST /files/v1 (direct multipart upload), use these endpoints to poll its status, download its converted results, or delete it.

Endpoint	Description
GET /files/v1/{file_id}	Poll processing status
GET /files/v1/{file_id}.{ext}	Download a converted result in the requested format
DELETE /files/v1/{file_id}	Permanently remove a file and its results

---

GET /files/v1/{file_id}

GET api.mathpix.com/files/v1/{file_id}

Returns the file's status and processing progress.

Poll until status is "completed" (or "error").

Example

cURL

curl -H 'app_key: APP_KEY' \
  https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d

Python

import requests
r = requests.get(
    "https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d",
    headers={"app_key": "APP_KEY"},
)
body = r.json()
print(body["status"], body["percent_done"])

JavaScript / TypeScript

const response = await fetch(
  "https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d",
  { headers: { app_key: "APP_KEY" } },
);
const file = await response.json();
console.log(file.status, file.percent_done);

Go

req, _ := http.NewRequest("GET", "https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d", nil)
req.Header.Set("app_key", "APP_KEY")
resp, _ := http.DefaultClient.Do(req)
defer resp.Body.Close()
result, _ := io.ReadAll(resp.Body)
fmt.Println(string(result))

Java

HttpClient client = HttpClient.newHttpClient();
HttpRequest request = HttpRequest.newBuilder()
    .uri(URI.create("https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d"))
    .header("app_key", "APP_KEY")
    .GET()
    .build();
HttpResponse<String> response = client.send(request, HttpResponse.BodyHandlers.ofString());
System.out.println(response.body());

Example response

{
  "percent_done": 100.0,
  "formats": {
    "md": "completed",
    "docx": "completed"
  },
  "custom_id": "contract-001",
  "num_pages": 30,
  "destination_uri": null,
  "destination_basename": null,
  "filename": "b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d.pdf",
  "format_primary": "mmd",
  "file_id": "b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d",
  "num_pages_completed": 30,
  "status": "completed"
}

Response body

file_id string

The file's identifier.

status string

Overall processing status; one of the status values below.

percent_done float

Progress from 0.0 to 100.0.

num_pages int

Total pages detected in the document. 0 until the page split runs.

num_pages_completed int

Pages that have finished OCR.

formats object

Per-requested-format conversion status; see Per-format conversion status below.

format_primary string

Always mmd.

filename string

The display name supplied at submit, or <file_id>.pdf when none was.

custom_id string (optional)

Echoed back when supplied at submit; null otherwise.

destination_uri string (optional)

The result destination supplied at submit; null when results stay in Mathpix storage.

destination_basename string (optional)

The output basename supplied at submit; null when defaulted.

Status values

Status	Meaning
pending	File registered, queued for processing.
split	Pages extracted, OCR and conversion in progress (poll percent_done).
completed	All processing finished; results available via download.
error	Processing failed; see the error fields on the response.

Per-format conversion status

The formats map carries one entry per format you requested via conversion_formats, each with its own conversion status (received / loaded / processing / completed / error). A requested format is absent from the map until its conversion starts; treat a missing entry the same as a not-yet-completed one.

Conversions complete independently of the top-level status and can lag behind it: a file can be completed overall while an individual format is still processing.

note

Poll formats.{ext} before downloading that extension; a download of a format that isn't yet completed returns 404 format_not_ready.

Error fields

When status is "error", the response carries the same error + error_info pair used by Files API request errors, alongside the usual fields:

Example response (status: error)

{
  "percent_done": 0.0,
  "formats": {},
  "error_info": {
    "id": "data_source_not_found",
    "message": "No data source registered for source"
  },
  "custom_id": "contract-001",
  "num_pages": 0,
  "destination_uri": null,
  "destination_basename": null,
  "filename": "b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d.pdf",
  "format_primary": "mmd",
  "file_id": "b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d",
  "num_pages_completed": 0,
  "error": "data_source_not_found",
  "status": "error"
}

error is a stable, machine-readable code (error_info.id duplicates it for v3-parser compatibility); see the error reference for the full list.

note

Remote-source fetching happens asynchronously, so source problems surface here, not on the original submit call. Common values: data_source_not_found (no data source registered for the bucket), data_source_access_denied (the bucket's grant isn't set up), and content_too_large.

---

GET /files/v1/{file_id}.{ext}

GET api.mathpix.com/files/v1/{file_id}.{ext}

Download a converted result.

The MMD format is always produced; other formats produce only when requested via conversion_formats on the original submission.

Supported extensions

mmd, md, md.zip, mmd.zip, docx, pptx, xlsx, html, html.zip, tex.zip, latex.pdf, pdf, lines.json, lines.mmd.json.

See Supported formats for the full list with descriptions.

Example

cURL

curl -H 'app_key: APP_KEY' \
  -o output.docx \
  https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d.docx

Python

import requests
r = requests.get(
    "https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d.docx",
    headers={"app_key": "APP_KEY"},
)
if r.status_code == 200:
    with open("output.docx", "wb") as f:
        f.write(r.content)
elif r.status_code == 404 and r.json().get("error") == "format_not_ready":
    print("Format still converting; retry later.")

JavaScript / TypeScript

import { writeFile } from "node:fs/promises";
const response = await fetch(
  "https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d.docx",
  { headers: { app_key: "APP_KEY" } },
);
if (response.ok) {
  await writeFile("output.docx", Buffer.from(await response.arrayBuffer()));
}

Go

req, _ := http.NewRequest("GET", "https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d.docx", nil)
req.Header.Set("app_key", "APP_KEY")
resp, _ := http.DefaultClient.Do(req)
defer resp.Body.Close()
out, _ := os.Create("output.docx")
defer out.Close()
io.Copy(out, resp.Body)

Java

HttpClient client = HttpClient.newHttpClient();
HttpRequest request = HttpRequest.newBuilder()
    .uri(URI.create("https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d.docx"))
    .header("app_key", "APP_KEY")
    .GET()
    .build();
client.send(request, HttpResponse.BodyHandlers.ofFile(Path.of("output.docx")));

Response headers

HTTP/1.1 200 OK
Content-Type: application/vnd.openxmlformats-officedocument.wordprocessingml.document
Content-Disposition: attachment; filename="b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d.docx"

The Content-Type matches the requested extension (for example text/plain for .mmd); the Content-Disposition filename is <basename>.<ext>.

Errors

Code	HTTP	When it fires
format_not_ready	404	Format is still converting (formats.{ext} is absent or not yet completed). The error body carries the file's current status. Retry after a short delay.
unsupported_format	415	Extension wasn't requested via conversion_formats on the original submission, or isn't a supported output format.
not_found	404	file_id doesn't exist (or was deleted).

note

lines.json and lines.mmd.json are available once the primary mmd format completes.

---

DELETE /files/v1/{file_id}

DELETE api.mathpix.com/files/v1/{file_id}

Permanently remove a file and its results from Mathpix-owned storage.

Files are auto-deleted on a per-artifact schedule (source and page images after 30 days, text outputs after 90 days; see Data retention). Call this to remove sooner.

Example

cURL

curl -X DELETE -H 'app_key: APP_KEY' \
  https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d

Python

import requests
r = requests.delete(
    "https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d",
    headers={"app_key": "APP_KEY"},
)
print(r.json())  # {"file_id": "...", "status": "deleted"}

JavaScript / TypeScript

const response = await fetch(
  "https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d",
  { method: "DELETE", headers: { app_key: "APP_KEY" } },
);
const { file_id, status } = await response.json();

Go

req, _ := http.NewRequest("DELETE", "https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d", nil)
req.Header.Set("app_key", "APP_KEY")
resp, _ := http.DefaultClient.Do(req)
defer resp.Body.Close()
result, _ := io.ReadAll(resp.Body)
fmt.Println(string(result)) // {"file_id": "...", "status": "deleted"}

Java

HttpClient client = HttpClient.newHttpClient();
HttpRequest request = HttpRequest.newBuilder()
    .uri(URI.create("https://api.mathpix.com/files/v1/b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d"))
    .header("app_key", "APP_KEY")
    .DELETE()
    .build();
HttpResponse<String> response = client.send(request, HttpResponse.BodyHandlers.ofString());
System.out.println(response.body());

Example response

{
  "file_id": "b1c9c3a8-55e4-4a09-b7d0-218ba5de4c4d",
  "status": "deleted"
}

Response body

file_id string

The deleted file's identifier.

status string

Always "deleted" on success.

Behavior

• Only terminal files can be deleted. A file still being processed (pending / split) cannot be deleted; DELETE returns 409 conflict. Wait for completed or error, then delete.
• Idempotent. Calling DELETE on an already-deleted file returns the same 200 / status: deleted body, not 404.
• Mathpix-owned storage only. Results delivered to a customer-owned bucket via destination_uri are not affected; those live under your bucket's own lifecycle policy. Mathpix never deletes from customer-owned buckets.
• Billing counters preserved. Per-month page and file counts that drive billing are never decremented. Deleting a file does not credit your account.
• Job counters preserved. A file's job remains intact; file_count / files_completed / files_errored on the parent job are not adjusted.

Errors

Code	HTTP	When it fires
not_found	404	file_id doesn't resolve to any row (and has never existed).
conflict	409	file_id exists but is still processing (pending / split); not yet deletable.
forbidden	403	file_id exists but is owned by a different group.

---

See also

• POST /files/v1/uri: submit a single file.
• POST /files/v1/jobs: submit a batch.
• Data retention: automatic cleanup schedule per artifact type.
• Supported formats: full list of conversion outputs.