Parsio

@Parsio list all documents in my invoices mailbox from the last 7 days and show me which ones failed parsing

Found 23 documents in the 'invoices' mailbox from the past week. 19 parsed successfully. 4 failed: two due to unexpected PDF layouts (documents #8472, #8491), one missing required fields (#8503), and one timeout during OCR (#8517). The failed documents are marked 'error' status and can be reparsed after template adjustments or manual review.

This example shows list-and-filter across document status. Parsio returns paginated results; the MCP handles iteration automatically but large mailboxes may take multiple API calls. Filtering by date and status requires the mailbox ID, which the MCP resolves from the mailbox name you provide.

@Parsio parse this email body as a new document in my 'customer-orders' mailbox and wait for the extracted fields

Document created (ID: doc_9x4k2m). Parsing completed in 3.2 seconds. Extracted fields: order_number='ORD-2847', customer_email='sarah@example.com', total_amount='$342.50', line_items=[3 products]. This operation consumed 1 parsing credit from your account balance (47 credits remaining this month).

This write operation submits text to Parsio's parser and waits for results, which can take 2–10 seconds depending on document complexity. Each parse consumes account credits. If the mailbox uses a GPT-powered parser, costs and latency may be higher. Do not retry failed parses without checking your credit balance and parser configuration first.

@Parsio get the parser configuration for my 'receipts' mailbox, then summarize what fields it extracts and whether it uses AI or template rules

The 'receipts' mailbox (ID: mbx_7h3p) uses a GPT-4-powered parser with 8 extraction fields: merchant_name, transaction_date, total, tax, payment_method, receipt_number, line_items (array), and notes. No template rules are configured—this mailbox relies entirely on AI inference. Average parse time: 4.1 seconds. The mailbox has processed 1,847 documents with a 94% success rate.

This synthesis example pairs mailbox metadata with the AI's ability to interpret parser type and field schema. Parsio supports both template-based and GPT-powered parsing; understanding which mode a mailbox uses helps predict cost, speed, and accuracy. The MCP fetches configuration but does not modify parser settings—that requires the Parsio dashboard.

When Parsio wins for inbound support ticket extraction

A 6-person support team receives 200+ emails daily—refund requests, bug reports, feature asks—all forwarded to a shared Parsio mailbox. The team uses List Documents each morning to pull the overnight batch, then Get Document to extract structured fields (customer ID, issue type, urgency flag) into their ticketing system. This works because Parsio's template parser handles the repetitive structure of support emails without burning LLM credits on every message. The threshold: if your emails vary wildly in format (mixed languages, attachments with no text body), the template approach breaks down and you'll spend more time tweaking parsers than triaging tickets. For teams with predictable inbound patterns, Parsio turns email chaos into queryable JSON in under 10 minutes of setup.

Why Parsio fits recurring document workflows, not ad-hoc pulls

A 3-person finance team at a SaaS company closes books on the 5th of each month. Vendors email invoices as PDFs or HTML throughout the prior month; the team forwards them to a Parsio mailbox configured to extract vendor name, amount, date, and line items. On close day, they run List Documents filtered by date range, then Parse Text Document for any stragglers that didn't auto-parse. The win: Parsio's mailbox model means invoices flow in passively all month, no manual uploads. The catch: if you need to extract data from a one-off contract PDF right now, Parsio's async queue (Reparse Document can take minutes) feels slow compared to a synchronous OCR API. This MCP shines when document ingestion is a recurring background task, not a blocking step in a live conversation.

When Parsio's field mapping beats manual CSV wrangling

A 4-person growth team runs weekly webinars. Registration confirmations arrive as HTML emails with attendee details buried in tables. They forward confirmations to a Parsio mailbox, use List Table Fields to confirm the parser extracts name, company, title, and LinkedIn URL, then pull the batch with List Documents after each event. The structured output feeds directly into their CRM via Zapier. This setup works because the confirmation template is stable week-to-week, so the parser rarely needs tuning. The boundary: if your lead source changes format every campaign (different event platforms, custom landing pages), you'll spend more time in Get Mailbox reconfiguring parsers than you save on manual data entry. For teams with a small set of repeating lead sources, Parsio turns email into CRM-ready records without writing a single regex.