Mathpix
Mathpix extracts text, mathematics, tables, and other structured content from images and documents and converts results into editable formats.
Verdict
Common use cases
- Extract equations from lecture slide PDFs
- Convert handwritten meeting notes to Markdown
- Parse formulas in scanned research papers
- Turn whiteboard photos into editable LaTeX
- Batch-process invoice images into structured text
Integration
- Vendor
- Mathpix
- Category
- docs
- Auth
- API_KEY
- Tools
- 29
- Composio slug
mathpix
Tools
- Convert Mathpix Markdown
Submit Mathpix Markdown, wait for one requested format to finish, then return the conversion ID and downloadable output file. The remote conversion remains if waiting fails or times out.
- Create App Token
Mint a short-lived client OCR credential, optionally with a live ink strokes session. Requires explicit acknowledgement because the response exposes the new secret exactly once to the caller.
- Delete Conversiondestructive
Permanently delete one Mathpix Markdown conversion's hosted output data after explicit confirmation. The response preserves the status immediately before deletion; Mathpix may continue to report that status after the output data has been de
- Delete Documentdestructive
Permanently delete one Mathpix document source and its hosted outputs after explicit confirmation. The response reports the status immediately before deletion, not a deleted status marker, and cached downloads may remain reachable for up to
- Delete Filedestructive
Permanently delete one terminal Files API file and its Mathpix-hosted results. This does not remove customer-owned destination objects, reverse billing, or delete a file that is still processing.
- Finalize File Job
Permanently close one Files API job to new file additions and arm its completion webhook. Already accepted files may continue processing. Requires explicit confirmation; repeating the same finalization is provider-idempotent.
- Get App Token
Get expiry and ownership metadata for a specific short-lived Mathpix app token. The secret token is used only in the request path and is omitted from the returned object and provider-error diagnostics.
- Get Conversion Result
Wait for an existing Mathpix Markdown conversion and requested format to finish, then return its status and downloadable output file without creating a conversion.
- Get Document Result
Wait for an existing Mathpix document job and requested format to finish, then return its status and downloadable output file without creating a new job.
- Get File Job
Get aggregate status, file and page counters, and timestamps for one Files API job. A completed job can still contain errored files; inspect files_errored and has_file_errors.
- Get File Result
Wait for an existing Files API file and selected format to finish, then return its status and downloadable output without creating a file.
- Get Job File by Custom ID
Get the complete Files API processing status for one job member using the exact caller-defined custom ID assigned at submission.
- Get OCR Usage
Return one page of aggregated Mathpix OCR usage for a date range, optionally grouped by usage type, request configuration, or application.
- Get Storage Onboarding Identities
Get the Mathpix AWS, Azure, and GCP identities needed to grant cloud-storage access. Requires explicit acknowledgement because the first call may lazily create the account's immutable external ID.
- List Conversions
List one page of retained Mathpix Markdown conversion records, optionally filtered by date, and return IDs usable by conversion result or deletion tools.
- List Data Sources
List cloud-storage data sources registered with Mathpix, returning only allowlisted non-secret metadata.
- List Documents
List retained document OCR jobs for the connected account, optionally filtered by date or document ID, and return one page.
- List File Job Files
List one cursor-paginated page of files in a Mathpix Files API job, optionally filtered by terminal status.
- List File Jobs
List one cursor-paginated page of Files API jobs newest first, optionally bounded by inclusive UTC dates.
- OCR Document
Submit a PDF, office document, or ebook URL, wait for OCR and the requested conversion to finish, then return its Mathpix job ID and downloadable output file. Mathpix retains the remote job after success; call Delete Document when the hoste
- OCR Image
Extract text, math, tables, diagrams, or chemistry from one image URL or base64 data URI and return the requested OCR representations. Mathpix may retain the submitted image and result, and provides no image deletion endpoint.
- Process Equation (Deprecated)
Compatibility-only access to Mathpix's deprecated equation-image OCR endpoint. Use MATHPIX_OCR_IMAGE for all new image OCR work. This paid operation may retain the submitted image and result in Mathpix history.
- Process File URI
Submit one public or registered-cloud document URI to the paid Files API, wait for OCR and one requested output, and return the retained file ID and downloadable result. Supply an idempotency key for safely retryable standalone submissions.
- Process Image Batch
Submit a paid asynchronous batch of 1-100 named public image URLs, wait until every key has a semantically successful OCR result, and return the retained batch ID and results. Explicit confirmation is required because Mathpix provides no ba
- Recognize Strokes
Recognize handwritten math or text from aligned x/y digital-ink stroke arrays and return OCR representations. This paid call may retain the result.
- Revoke App Tokendestructive
Irreversibly revoke one short-lived app token immediately. Existing clients using the credential will stop authenticating. Requires explicit confirmation.
- Search OCR Results
Search retained image and stroke OCR history by date, identifiers, content, tags, or detected content and return one page of results.
- Submit File Job
Submit 1 to 200,000 remote documents as one paid asynchronous Mathpix Files job. This creates retained processing state and may read or write external storage or deliver callbacks. Acceptance defers per-file validation, so use job status an
- Upload and Process File
Upload one local document to the paid Mathpix Files API, wait until OCR and one requested output are ready, then return the retained Mathpix file ID and a downloadable result. The operation can consume page quota and does not delete the rem
Setup
Setup guide
- 11. Open your Switchy workspace settings and navigate to the Integrations tab. 2. Click 'Add MCP' and select Mathpix from the catalog. 3. Log into your Mathpix account at mathpix.com, navigate to the API section in your dashboard, and generate a new API key. 4. Paste the API key into Switchy's connection dialog and click 'Connect'. 5. Switchy will verify the key by pinging the Mathpix API; you'll see a green checkmark when authentication succeeds. 6. Open any Project in your workspace and type '@Mathpix' in the chat to confirm the integration appears in the autocomplete list. 7. Test the connection by uploading a screenshot of an equation and asking '@Mathpix convert this to LaTeX' — you should receive formatted output within seconds. 8. Check your Mathpix dashboard to monitor quota usage; each OCR request and conversion counts toward your plan limit.
What teammates see: by default, memories from Mathpix 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
Extract Slide Equations
@Mathpix extract all equations from the attached lecture slides and return them as LaTeX. Group by slide number.
Handwriting to Markdown
@Mathpix convert this handwritten page to Markdown. Preserve the heading hierarchy and bullet points I drew.
Parse Research Paper
@Mathpix parse the attached research paper PDF. Return the abstract, all equations, and figure captions as structured text.
Whiteboard to LaTeX
@Mathpix convert this whiteboard photo to LaTeX. Include every step of the derivation and label each equation.
Batch Invoice OCR
@Mathpix extract line items, totals, and vendor names from these five invoice images. Return a table I can paste into a spreadsheet.
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@Mathpix extract all equations and tables from the PDF in our shared drive and return them as LaTeX
This example illustrates Mathpix converting a technical document. The MCP would return extracted content in Mathpix Markdown format, with equations rendered as LaTeX (e.g., `\int_{0}^{\infty} e^{-x^2} dx = \frac{\sqrt{\pi}}{2}`) and tables preserved as structured markdown. A conversion ID would be provided to retrieve the full output file. The actual response depends on the document's complexity and your account's processing queue.
Showcases Mathpix's core OCR strength for mathematical and scientific content. Requires a valid API key with sufficient credits. Large documents may take 30+ seconds to process; the MCP polls for completion but will time out on very long conversions, leaving the job running remotely for later retrieval.
@Mathpix convert the handwritten notes image I uploaded to editable DOCX format
This example shows the MCP submitting Mathpix Markdown for conversion to DOCX. After processing, you'd receive a conversion ID and a download URL for the Word document. Handwritten equations would be converted to equation objects; prose would become editable text. The conversion remains hosted on Mathpix servers until you explicitly delete it or it expires per your account policy.
Demonstrates the two-step workflow: OCR to Mathpix Markdown, then format conversion. Handwriting recognition quality varies with penmanship and image resolution. The Create/Delete Conversion tools require explicit confirmation to prevent accidental data loss or unexpected billing.
@Mathpix compare the equations in last week's draft paper against the final version and summarize what changed mathematically
This example pairs Mathpix OCR with AI reasoning. The MCP would extract LaTeX from both documents, then the AI would diff the equation sets and explain changes (e.g., 'The diffusion coefficient changed from D=0.5 to D=0.8 in equation 3; boundary condition in equation 7 switched from Dirichlet to Neumann'). This workflow requires two separate OCR calls and relies on the AI to interpret mathematical differences, not Mathpix itself.
Highlights synthesis use-case where Mathpix handles extraction and the AI handles analysis. Each document consumes API credits independently. Rate limits apply if processing many files in quick succession; consider batching or using the Files API job workflow for bulk operations.
Use-case deep-dives
When Mathpix wins for converting legacy academic PDFs
A university research group inherits 200 scanned journal articles from a retired professor and needs searchable markdown for their Notion wiki. Mathpix is the right call here because its OCR handles equations, tables, and multi-column layouts that break standard PDF extractors. The team uses the Convert Mathpix Markdown tool to batch-process files overnight, then the Get Conversion Result tool to pull finished outputs into their workspace. The API_KEY auth means one admin can provision access without per-user licensing. The threshold: if your PDFs are already text-layer and don't have math notation, a simpler extractor is faster and cheaper. But for scanned academic content with LaTeX-heavy equations, Mathpix saves 40 hours of manual retyping and the output pastes cleanly into any markdown editor.
When this MCP handles real-time sketch digitization
A 5-person product team runs weekly remote design sprints where participants sketch wireframes on tablets and need those drawings converted to editable text and diagrams within 10 minutes. Mathpix's ink strokes session (via Create App Token) lets the team stream handwriting in real time and get structured markdown back before the meeting ends. The Delete File and Finalize File Job tools keep the workspace tidy by removing intermediate uploads once the sprint doc is finalized. This works when sketches include annotations, arrows, and rough math that need OCR, not just vector shapes. The boundary: if your team draws in Figma or Miro already, you don't need OCR. But for hybrid teams mixing paper notebooks, tablet styluses, and photo uploads, Mathpix turns messy input into clean documentation without a transcription step.
When Mathpix extracts text from user-submitted screenshots
A 12-person SaaS support team receives 80 tickets daily with screenshots of error messages, invoices, and handwritten notes. Mathpix's Files API (Delete Document, Get Conversion Result) lets agents upload images and get searchable text back in under 5 seconds, so they can route tickets by keyword without squinting at JPEGs. The 29-tool count includes batch operations and webhook finalization, which matters when processing spikes during product launches. API_KEY auth means the support lead provisions one credential and tracks usage centrally. The trade-off: if screenshots are already high-contrast text (like terminal output), a lighter OCR service costs less per call. But when tickets mix handwriting, tables, and low-resolution phone photos, Mathpix's accuracy justifies the price and the team closes tickets 20% faster because they find answers in the first search.
Frequently asked
What does the Mathpix MCP do in Switchy?
It connects Switchy to Mathpix's OCR and document conversion APIs. Your team can extract LaTeX from handwritten equations, convert PDFs to Markdown, and process scientific documents without leaving the workspace. The MCP wraps all 29 Mathpix endpoints, so you can automate workflows that previously required copy-pasting between Mathpix's web UI and your tools.
Do I need a Mathpix account to use this MCP?
Yes. You need an active Mathpix subscription and an API key from your Mathpix dashboard. The MCP authenticates with that key, so whoever connects it must have access to your organisation's Mathpix credentials. Free-tier Mathpix accounts work, but conversion limits and OCR quotas follow your Mathpix plan, not Switchy's.
Can the MCP process images I upload directly to Switchy?
No. The MCP sends requests to Mathpix's hosted API; it doesn't run OCR locally. You reference files already uploaded to Mathpix or pass public URLs. If your image lives in Switchy storage, you'll need to expose a temporary signed URL or upload it to Mathpix first using the Files API tools this MCP includes.
How is this different from using Mathpix's web app?
The web app is faster for one-off conversions. The MCP is better when you're chaining Mathpix into multi-step workflows—extract equations from a batch of PDFs, feed the LaTeX into another tool, log results to Notion. You also get programmatic access to app tokens and job management, which the web UI doesn't expose.
Who on the team should connect this integration?
Whoever holds your Mathpix API key and understands your OCR budget. The Create App Token and Delete tools can mint credentials and wipe hosted data, so treat the connection like admin access. If multiple people need OCR, connect it once under a shared service account rather than each person burning their own Mathpix quota.