Private document AI

Extract documents with AI and control your processing route

के साथ शुरू

Processing settings

डायनामाइट डॉक्स

इसके लिए काम करें

प्रसंस्करण नियंत्रण

Choose the processing route · overviewProcessing settings. Check provider terms before processing.. File storage: Browser-only or encrypted R2-backed library. Provider choice: Hosted route, own key or local Ollama. Training policy: Sensitive and no-training-only routing rules Illustrative diagram, not extracted customer data.स्रोतProcessing settingsFile storageBrowser-only or encrypted…Provider choiceHosted route, own key or…Training policySensitive and…समीक्षा हेतु आउटपुटChoose the processingrouteFile storage: Browser-only orencrypted R2-backed libraryProvider choice: Hostedroute, own key or local…Training policy: Sensitiveand no-training-only routing…नीचे पूरी नीति पढ़ेंCheck provider terms before processing.

उदाहरण स्रोत

Processing settings

File storage
Browser-only or encrypted R2-backed library
Provider choice
Hosted route, own key or local Ollama
Training policy
Sensitive and no-training-only routing rules

Check provider terms before processing.

स्थानीय ओलामामिस्ट्राल
Check provider terms before processing. Illustrative workflow.

Bring your own AI OCR or run local inference without third-party data retention.

Document privacy is a chain of choices, not one toggle. File storage, model routing, provider retention, account data and deletion all have separate boundaries. Dynamite Docs lets teams bring their own AI provider keys or run zero data egress workflows without vendor lock-in.

Use browser-only files for a quick trial, encrypted cloud storage when documents must persist, own-key routing when the provider account should stay under your control, or the local Ollama companion when model inference should run on your computer.

Try a private workflow. Free files stay browser-only. Every plan supports BYOK; BYOK on Pro and Scale is unlimited.

Choose the control before the upload

  • Browser-only files: Anonymous and Free uploads do not become R2-backed library files.
  • Encrypted cloud files: Starter, Pro, and Scale can store supported files for later work.
  • Provider-controlled inference: Every plan can connect encrypted provider keys.
  • Local model inference: Starter, Pro, and Scale can use the signed Ollama companion.

Controls that affect sensitive documents

Each control answers a different question. Check the route as a whole before treating it as suitable for financial, identity or legal records.

  • File storage: Browser-only or encrypted R2-backed library
  • Provider choice: Hosted route, own key or local Ollama
  • Training policy: Sensitive and no-training-only routing rules
  • Residency policy: EU-only policy on Pro and Scale
  • Retention: Delete files or use published retention controls
  • Audit evidence: Extraction and account events on eligible plans

फ़ाइल से लेकर समीक्षित परिणाम तक

  1. 1. Classify the document. Decide whether the file contains personal, financial, legal or regulated information.
  2. 2. Choose storage. Keep the file browser-only or use paid cloud storage when cross-device persistence is required.
  3. 3. Set the model rule. Use hosted routing, your provider key, a no-training policy or local-only inference.
  4. 4. Review and delete. Check the extraction, export only what is needed and remove stored files under your retention policy.

Provider and policy controls are visible

AI Settings shows available providers, keys and routing rules. A policy can narrow eligible models, but the selected provider terms and your account configuration still matter.

डायनामाइट डॉक्स लाइब्रेरी का वास्तविक आयात मेनू, जिसमें अपलोड और गूगल ड्राइव के विकल्प शामिल हैं।
वास्तविक एआई प्रोसेसिंग डायलॉग बॉक्स, जिसमें खोज योग्य मॉडल सूची और प्रदाता फ़िल्टर दिखाए गए हैं।
टेक्स्ट एडिटर में निकाले गए टेक्स्ट के बगल में एक खरीद आदेश।
टेबल एडिटर में संरचित विवरण के साथ एक खरीद आदेश।
टेबल एडिटर में निकाले गए लाइन आइटम के बगल में एक खरीद आदेश।
टेबल एडिटर के टोटल टैब के बगल में एक परचेज़ ऑर्डर है, जिसमें टैक्स और अंतिम टोटल दिखाया गया है।
निकाले गए दस्तावेज़ फ़ील्ड और विश्वास संकेतकों के साथ-साथ खरीद आदेश का पाठ।
फ़ाइल का प्रारूप, दायरा, फ़ाइल का नाम और फ़ोल्डर विकल्प सहित Google ड्राइव में निर्यात करने का वास्तविक संवाद बॉक्स।
Dynamite Docs AI settings with providers, encrypted keys and routing policies
The real AI configuration panel used to choose providers and routing rules.

Control is different at each layer

The table separates what runs in the browser, the Dynamite Docs service, an external provider account and the optional local companion.

LayerWhat it handlesWhat to verify
BrowserUpload handoff and local file stateDevice access and browser storage
डायनामाइट डॉक्सAccount, workspace, routing and optional cloud filesPlan, policy, deletion and retention
AI providerModel inference for routed documentsProvider terms, region and retention
Local companionModel inference through Ollama on your machineLocal host, model and hardware

Privacy controls do not replace document review

आत्मविश्वास

A private route can still return an uncertain value. Keep source review, validation and approval separate from the processing path.

शुद्धता

Provider choice affects capability as well as policy. A model that meets the routing rule may still struggle with handwriting, dense tables or poor scans.

विफलता की स्थिति

A restrictive policy may leave no eligible provider. Local inference also fails when the companion, Ollama or selected model is unavailable.

उत्पाद सीमा

The local companion moves model inference to the user's computer. It is not a self-hosted copy of the account, workspace, billing or token service.

Shared responsibility for private extraction

Dynamite Docs publishes the controls it operates. Customers still choose a plan, provider, policy, access model and retention schedule that fit their obligations.

Areaडायनामाइट डॉक्सCustomer
AccessAuthenticated workspace and capability gatesManage users, devices and shared exports
Provider keys[AES-256-GCM encrypted storage](/docs/document-security-and-data-residency) on every signed-in planChoose the provider account and rotate keys
FilesBrowser-only option and encrypted paid storageClassify, upload and delete the right documents
AI routingPolicy engine and eligible-provider filteringSelect policies and review provider terms
ResultsConfidence signals and editable tablesValidate values before operational use

Read the policy pages before a security review

The Security page documents encryption and provider controls. The Data Processing Agreement covers customer data processing, and the retention policy explains how stored records and deletion work.

Choose storage, retention, and provider settings that match your document requirements. Use the published controls and contractual documents when completing your security review.

Questions about private document ai

Is BYOK available on every plan?

Yes. Free and Starter own-key processing uses monthly PE. Pro and Scale BYOK is unlimited and does not use hosted PE.

Does local AI make the whole product self-hosted?

No. The companion performs model inference on your computer. Account, workspace and token issuance remain hosted services.

Can I prevent routing to training-permitted providers?

Yes. Sensitive Documents mode and no-training-only rules restrict eligible providers. Confirm the active rule and current provider terms before processing a sensitive file.

What is the difference between per-page OCR billing and BYOK unlimited extraction?

Metered OCR bills for provider usage under the vendor's pricing rules. BYOK routes model calls through your chosen provider account, while local Ollama runs inference on your machine. Pro and Scale do not charge Dynamite Docs PE for BYOK runs, but the external provider may still bill its own usage.

अगला दस्तावेज़ कार्य चुनें

संबंधित वर्कफ़्लो

Try a private workflow or compare processing, storage and integration limits.

के साथ शुरू

Processing settings

डायनामाइट डॉक्स

इसके लिए काम करें

प्रसंस्करण नियंत्रण

Choose the processing route · workflowProcessing settings. Classify the document → Choose storage → Set the model rule → Review and delete. Illustrative diagram, not extracted customer data.Processing settings1Classify the document2Choose storage3Set the model rule4Review and deleteप्रत्येक चरण में दस्तावेज़ और परिणाम आपस में जुड़े रहते हैं।
  1. 01Classify the document
  2. 02Choose storage
  3. 03Set the model rule
  4. 04Review and delete
स्थानीय ओलामामिस्ट्राल
Classify the document → Choose storage → Set the model rule → Review and delete Illustrative workflow.

के साथ शुरू

Processing settings

डायनामाइट डॉक्स

इसके लिए काम करें

प्रसंस्करण नियंत्रण

Choose the processing route · reviewProcessing settings. A private route can still return an uncertain value. Keep source review, validation and approval separate from the processing path.. Provider choice affects capability as well as policy. A model that meets the routing rule may still struggle with handwriting, dense tables or poor scans.. A restrictive policy may leave no eligible provider. Local inference also fails when the companion, Ollama or selected model is unavailable.. Illustrative diagram, not extracted customer data.इससे पहले कि आप इस पर भरोसा करेंपरिणामChoose the processing route1A private route can still return an uncertain value. Keepsource review, validation and approval separate from theprocessing path.2Provider choice affects capability as well as policy. Amodel that meets the routing rule may still struggle withhandwriting, dense tables or poor scans.3A restrictive policy may leave no eligible provider.Local inference also fails when the companion, Ollama orselected model is unavailable.Check provider terms before processing.
  1. 01A private route can still return an uncertain value. Keep source review, validation and approval separate from the processing path.
  2. 02Provider choice affects capability as well as policy. A model that meets the routing rule may still struggle with handwriting, dense tables or poor scans.
  3. 03A restrictive policy may leave no eligible provider. Local inference also fails when the companion, Ollama or selected model is unavailable.
स्थानीय ओलामामिस्ट्राल
A private route can still return an uncertain value. Keep source review, validation and approval separate from the processing path.. Provider choice affects capability as well as policy. A model that meets the routing rule may still struggle with handwriting, dense tables or poor scans.. A restrictive policy may leave no eligible provider. Local inference also fails when the companion, Ollama or selected model is unavailable. Illustrative workflow.
डायनामाइट दस्तावेज़ लोड हो रहे हैं... इस पेज को लोड होने में उम्मीद से ज़्यादा समय लग रहा है। कृपया पेज को पुनः लोड करें ।