Learn to extract, review, and export document data with AI

01. Start extracting document data with AI

Your first document, from zero to structured data

Dynamite Docs uses AI to read PDFs, scans, and images, propose fields and table rows, and place the result beside the source for review. Correct the values, then export the accepted data.

Sign in with your account for higher limits and cloud storage (on paid plans); anonymous visitors can try extraction with tighter, per-IP daily caps.

Drag and drop a document anywhere into the file panel, or press Ctrl+U. The file appears in the sidebar immediately.

Select the file and press Ctrl+Enter (or click Extract in the Data Studio) and the extracted table lands in the panel below.

Fix any wrong values directly in the table; corrections are saved automatically and reused as patterns for future documents.

Upload a supported file, select it, and press Ctrl+Enter or choose Extract. The result opens as an editable table, and saved corrections can guide later files with the same layout.

Read the getting started chapter.

02. Upload PDFs, images, and documents for AI extraction

What you can upload, and where your files live

Upload supported PDFs, images, Office files, OpenDocument files, email, and text files to Dynamite Docs. Free originals stay in browser storage, while paid plans can store eligible files in the cloud library.

Supported formats include PDFs (including scanned), images (JPG, PNG, and other raster formats), and text-based files. Word (.docx/.doc), Excel (.xlsx/.xls), CSV, PowerPoint, OpenDocument (ODT/ODS/ODP), HTML, RTF, and email files are extracted directly as text.

FREE uses browser-only storage (blob URLs); HOBBY adds 500 MB, PRO adds 2 GB, and ULTRA adds 10 GB of cloud (R2) storage synced across devices.

If cloud storage is exceeded, uploads fall back to browser-only.

On Pro and Ultra, the Email capture panel gives you a private forwarding address. Forwarded receipts or invoices land in your dashboard and are extracted automatically.

Persisted uploads are limited to 50 MB per file. Free and anonymous files stay browser-only; Starter, Pro, and Scale provide 500 MB, 2 GB, and 10 GB of cloud storage.

Read the uploading documents chapter.

03. Run document extractions using AI and review the results

How the infer → correct → remember loop works

Dynamite Docs proposes fields and table rows from each document. You compare them with the source, correct mistakes, and save confirmed corrections as patterns for similar layouts.

Extraction works in three steps: infer (the AI reads the document and infers which columns matter), correct (you review and edit anything wrong in Data Studio), and remember (every correction becomes a reusable pattern).

Select one or more files and press Ctrl+Enter, or use the Extract button. Results appear as an editable table. For whole batches, select files and use Add to job to build a named Extraction job (Hobby and above), then press Run job. Hobby permits 25 files per job, 10 active jobs, and unlimited runs. Each document is extracted on its own into a View Extractions folder named after the job.

Each page successfully processed with a hosted model costs 1 page-equivalent (PE). Failed extractions cost 0 PE.

BYOK is available on every plan. BYOK on Free and Hobby uses the monthly PE allowance; BYOK on Pro and Ultra is unlimited. A run counts as BYOK only when it uses your own provider API key.

Each plan caps hosted pages per document. BYOK has no per-document tier cap, though Free and Hobby BYOK still uses monthly PE and provider and file-size limits still apply.

Each successfully processed hosted or local no-AI page costs 1 page-equivalent. Failed pages cost 0, and own-key processing follows the allowance rules of the active plan.

Read the extracting data chapter.

04. Review and correct AI-extracted tables beside the source

Review, edit, save, and export your tables

Data Studio keeps the source document beside an editable table. Review confidence cues, correct cells, add or remove columns, and save the result before export.

The Data Studio panel shows the active extraction as an editable table. It has the same shape as a spreadsheet, but is built for documents.

Click any cell and type to edit values inline; edits are saved automatically.

Add or remove columns to shape the schema so it matches how you track your data.

Press Ctrl+S to commit the table, or Ctrl+Shift+S to export it as CSV, JSON, or XLSX. These downloads are available on Free after signup.

Edits are made directly in table cells. The reviewed column order, added fields, and corrected values carry into CSV, Excel, and JSON exports.

Read the data studio chapter.

05. Export reviewed AI extractions to Excel, CSV, and JSON

CSV, JSON, clipboard, and imports

Export the table you reviewed as Excel, CSV, or JSON, or push it to Google Sheets after connecting the account. The export uses the columns and corrected values visible in Data Studio.

Every extraction can be exported without changing the table: CSV opens in Excel, Sheets, and Numbers, while JSON is the raw structured output for pipelines and integrations.

Paste a CSV into the extractions panel to create a table without a document. Useful for combining data from other systems.

On Hobby and above, the Document Editor can save rich-text documents as Word (.docx) or PDF files.

Column order, edited values, and added columns all carry into the export exactly as you left them.

CSV is a flat table, Excel can preserve separate sheets and cell types, and JSON carries columns, rows, and document metadata. Check the file in its receiving system.

Read the exporting & sharing chapter.

06. Automate AI document extraction with APIs and webhooks

Pull structured data into your own stack

Use the /api/v1 endpoints to submit a document, retrieve an extraction, or download a result. Register a webhook when your system needs a signed extraction.completed notification.

The key-authenticated public API lives under /api/v1. A scoped key can list and export extractions, check usage, and submit a document for inference; file, settings, account, and billing changes still require a signed-in workspace session.

Export a completed extraction with GET /api/v1/extractions/:id/export?format=csv|json|xlsx, or submit a document with POST /api/v1/infer. Only the key's SHA-256 hash is stored.

Register up to 5 webhook endpoints per account; each extraction.completed POST carries an HMAC-SHA256 signature in the X-Dynamite-Signature header computed over the raw body.

Webhook deliveries retry with exponential backoff and jitter on network errors, 429s, and 5xx. Up to 3 attempts with a 10-second timeout.

Connect Google Sheets once and push any extraction with one click. A fresh tab per push, plus a Document Fields tab for metadata.

The REST API authenticates requests via SHA-256 hashed API tokens, while outbound webhooks deliver payload notifications with HMAC-SHA256 signatures across 3 retry attempts.

Read the api & webhooks chapter.

07. Organize document batches, tag records, and filter extraction history

Keep your documents organized

Use folders to group stored files and extractions, then add labels for another way to filter the work. A folder can also supply the files for a named batch extraction job.

Folders group your documents and extractions into a tree, and labels add an extra layer of categorization on top.

Create a folder with Ctrl+Shift+N or the new-folder button, then drag files in and out freely.

Right-click a folder to turn its documents into a batch Extraction job, then run it to extract everything in one pass.

Press Ctrl+F to jump to search and find any file by name.

Starter, Pro, and Scale support durable jobs. The active plan controls files per run and concurrent jobs, while the job record keeps each file and its status visible.

Read the folders & labels chapter.

08. Configure document extraction with your own LLM API keys

Hosted models, or bring your own key (BYOK)

Choose a hosted model, connect your own supported provider key, or run model inference through the local Ollama companion. Dynamite Docs handles document review and export; the selected model reads the document.

Dynamite Docs ships with hosted AI providers, including Mistral, Gemini, OpenAI, Anthropic, Groq, OpenRouter, and Cloudflare. Extraction routes to the best available provider automatically.

HOBBY, PRO, and ULTRA include premium hosted models; FREE uses hosted standard models.

BYOK is available on every plan. Connect any supported provider or OpenAI-compatible endpoint. Free and Hobby runs use monthly PE; Pro and Ultra runs are unlimited.

Install the Windows companion from Microsoft Store to connect your web workspace to local Ollama. Local inference is available on Hobby, Pro, and Ultra; use browser-only storage to keep source files on your computer.

Provider keys are encrypted with AES-256-GCM and used only to route your own extraction requests.

Dynamite Docs encrypts saved provider keys with AES-256-GCM. Own-key runs use monthly PE on Free and Starter and cost 0 PE on Pro and Scale.

Read the ai providers & keys chapter.

09. Understand security and data residency for AI extraction

Encryption, provider isolation, and training controls

Dynamite Docs uses TLS for network traffic and AES-256-GCM for saved provider keys. You can choose hosted processing, your own provider account, or local Ollama inference based on the document policy.

All traffic is encrypted in transit (TLS), stored documents and tables are encrypted at rest, and API keys are encrypted with AES-256-GCM before storage.

Dynamite Docs does not use customer documents to train foundation models. Hosted and own-key requests are subject to the selected provider's handling terms.

AI Settings exposes routing rules that lock extraction to the providers you trust: Sensitive Documents mode, no-training-only, free-models-only, and local-only.

Connecting your own provider means your data never touches a model you did not choose.

Dynamite Docs does not use customer documents to train models. The selected AI provider may receive document content, so review its retention and training policy before processing sensitive files.

Read the security & privacy chapter.

10. Understand AI extraction plans and processing limits

What each plan includes

The plan table lists monthly page-equivalents, hosted pages per document, cloud storage, API calls, job limits, and own-key processing rules for Free, Starter, Pro, and Scale.

Free is $0 or ₹0 monthly, with 25 PE, 5 hosted pages per document and Browser-local files. Hosted and BYOK processing share the monthly allowance. Starter 100 is $13 or ₹259 monthly, with 100 PE, 50 hosted pages per document and 500 MB cloud storage. Hosted and BYOK processing share the monthly allowance. Starter 500 is $34 or ₹879 monthly, with 500 PE, 50 hosted pages per document and 500 MB cloud storage. Hosted and BYOK processing share the monthly allowance. Pro 1K is $59 or ₹1,299 monthly, with 1,000 PE, 500 hosted pages per document and 2 GB cloud storage. Unlimited BYOK. Pro 2K is $104 or ₹2,199 monthly, with 2,000 PE, 500 hosted pages per document and 2 GB cloud storage. Unlimited BYOK. Scale 5K is $259 or ₹5,199 monthly, with 5,000 PE, 1,000 hosted pages per document and 10 GB cloud storage. Unlimited BYOK. Scale 10K is $439 or ₹8,699 monthly, with 10,000 PE, 1,000 hosted pages per document and 10 GB cloud storage. Unlimited BYOK. Scale 15K is $609 or ₹12,199 monthly, with 15,000 PE, 1,000 hosted pages per document and 10 GB cloud storage. Unlimited BYOK.

Hosted extraction stops at the monthly PE allowance. There is no surprise overage billing on any plan.

BYOK provider models are available on every plan. Own-key processing is unlimited on Pro and Ultra.

Anonymous (not signed in) usage is limited to a daily anti-abuse cap.

Free includes 25 PE per month. Scale includes 5,000 PE per month, and Pro and Scale own-key runs cost 0 PE. Monthly caps stop processing instead of adding automatic overage charges.

Read the plans & limits chapter.

11. Keyboard commands for fast cell editing, row navigation, and review

Move fast, stay in the flow

Use keyboard shortcuts to upload files, start extraction, save table edits, export results, create folders, search the library, and select or delete files.

Ctrl+Enter extracts the selected document(s); Ctrl+S saves the active table; Ctrl+Shift+S exports CSV, JSON, or XLSX after signup.

Ctrl+U uploads a file, Ctrl+Shift+N creates a new folder, and Ctrl+F focuses search.

Ctrl+A selects all files, Delete removes the selected files, and ? shows the shortcut help toast.

Ctrl+Enter starts extraction, Ctrl+S saves the active table, Ctrl+Shift+S opens export, Ctrl+U uploads files, and Ctrl+F focuses search.

Read the keyboard shortcuts chapter.

12. Fix missed rows and scan issues in AI document extraction

Common issues and quick fixes

Use this guide when an upload, extraction, export, or local Ollama connection fails. Start with the error shown in the app, then check the file, quota, provider, or local service involved.

If extraction fails or returns empty, retry once, then check the document: password-protected PDFs, dense or unusual layouts, and low-quality scans are the usual culprits. Re-scan at a higher DPI or flatten the file.

For local Ollama issues, verify Ollama is serving on port 11434 and the Microsoft Store companion is active on port 8756, then reconnect from AI Settings.

If you hit the monthly PE limit on Free or Hobby, hosted and BYOK extraction pauses until the allowance resets. Upgrade from the footer for unlimited Pro or Ultra BYOK processing.

Browser-only Free storage is local to the browser where it was created; Hobby, Pro, and Ultra files use their cloud allowance and follow you across devices.

If an upload falls back to browser-only storage, your cloud storage is full. Delete files to free space immediately or upgrade for more capacity.

For an empty or failed result, check password protection, missing pages, scan quality, and the selected model. Local Ollama also requires Ollama on port 11434 and the companion on port 8756.

Read the troubleshooting chapter.

13. Answers to common questions about AI document extraction

Straight answers to the questions we hear most

Use the FAQ to check whether a file needs a supplier template, how page-equivalents are charged, where files are stored, and which processing options are available.

No templates or training. Upload a document and Dynamite Docs infers the schema automatically; missed fields are added during review and remembered for the next file with the same layout.

Every field carries a confidence score; fixing an inline correction becomes a pattern so the next batch is more accurate.

Each successfully processed hosted page costs 1 page-equivalent (PE). Free and Hobby BYOK pages use the same allowance. Failed extractions and Pro or Ultra BYOK runs cost 0 PE.

Dynamite Docs does not use customer documents to train foundation models. Check the selected provider's handling terms and active routing policies.

A supplier template is not required for the first run. Dynamite Docs proposes the fields and rows, and a confirmed correction can become a reusable pattern for a similar layout.

Read the faq chapter.

14. Get in touch with the product team for technical questions and custom schemas

Talk to a human

Open the support desk from a signed-in workspace or email info@dynamitedocs.com. Include the affected file type, the action you took, and the exact error or unexpected result.

If the guide does not solve the issue, open the support desk from the app or email info@dynamitedocs.com with the relevant document details.

Include the file format, error message, or unexpected output so we can trace the extraction run directly.

Signed-in users can keep account-linked support conversations inside the app. Email remains available for product, billing, integration, and account questions.

Read the contact support chapter.