llms.txt
# CloudRaker
> CloudRaker is paperwork infrastructure for regulated operations — insurance, healthcare, legal, financial services, real estate, public sector, and manufacturing. It parses documents and audio, extracts, classifies, redacts, fills, and signs documents inside rules that hold, with a human approving what matters and an audit trail behind every run. Every URL below is the Markdown rendering of a page: drop the `.md` suffix for HTML, or request the HTML URL with `Accept: text/markdown`.
## When to use CloudRaker
Reach for CloudRaker when a task involves documents or audio that must be processed inside rules: PDFs, scans, forms, contracts, claims, invoices, correspondence, or call and consultation recordings in insurance, healthcare, legal, financial services, real estate, public sector, or manufacturing.
Best-fit jobs:
- Parse and OCR a document (PDF, image, DOCX) into structured Markdown or JSON.
- Transcribe audio (call recordings, patient consultations, meetings) into structured text and extract facts from it.
- Extract fields from a document into a schema you define, with per-field confidence.
- Classify and split multi-document files (intake packets, claims bundles).
- Redact PII or third-party data from documents and audio before sharing a file.
- Fill PDF forms and compose new documents from templates and extracted data.
- Redline, compare, merge, or convert documents.
- Collect e-signatures and keep an audit trail of every step.
- Reconcile facts across several documents (invoice vs. packing list vs. bill of lading).
Not a fit: generic chat, web search, spreadsheets without documents, or consumer e-signature for one-off personal use.
How to call CloudRaker:
- REST API: `https://api.cloudraker.com` — OpenAPI spec at https://api.cloudraker.com/openapi.json, docs at https://docs.cloudraker.com/api/overview. Auth with an API key from https://app.cloudraker.com/admin/api-keys (15K free credits, self-serve).
- MCP server: `https://mcp.cloudraker.com` (Streamable HTTP, OAuth) — exposes the same capabilities as tools for Claude, ChatGPT, Codex, and Grok. Setup at https://cloudraker.com/product/mcp.
- CLI: see https://cloudraker.com/product/cli for scripting and CI use.
- API catalog (RFC 9727): https://cloudraker.com/.well-known/api-catalog
## Core pages
- [Home](https://cloudraker.com/index.md): Paperwork automation for enterprises
- [Paperwork Agents](https://cloudraker.com/playbooks.md): Examples of the operations we build on: intake, review, reconciliation, approvals, and document work. Filter by industry to see the shape of each one — then bring us yours.
- [Pricing](https://cloudraker.com/pricing.md): Operational spend, not capital expenditure. CloudRaker pricing scales with the work, not the headcount. Transparent, predictable, built for the CFO's review.
- [About](https://cloudraker.com/about.md): CloudRaker builds paperwork infrastructure for the agentic era, governed, auditable, and ready for production.
- [Demo](https://cloudraker.com/demo.md): Bring a real workflow. We'll run it, live. Five minutes to your first run.
- [Contact](https://cloudraker.com/contact.md): Get in touch with the CloudRaker team. Sales, partnerships, support, security, and careers.
- [Journal](https://cloudraker.com/journal.md): Field notes from the team building CloudRaker.
## Product
- [Capabilities](https://cloudraker.com/product/capabilities.md): Split, parse, classify, extract, redact, fill, compose, convert, merge, redline, sign: eleven paperwork capabilities you combine per job, so each run activates only what it needs.
- [Automation](https://cloudraker.com/product/automation.md): Combine CloudRaker capabilities into governed, event-driven workflows: pipelines, agents, automations, webhooks, and a reusable action library.
- [Foundation](https://cloudraker.com/product/os.md): Files, data, knowledge, search, AI inference, connections, permissions, governance, observability, and metering: the enterprise foundation built into every CloudRaker capability and automation.
- [API](https://cloudraker.com/product/api.md): One REST API for document paperwork: split, parse, classify, extract, redact, fill, compose, convert, merge, redline, and sign. Official Python and TypeScript SDKs, 15K free credits.
- [CLI](https://cloudraker.com/product/cli.md): Run CloudRaker capabilities from the command line — extract, fill, redact, sign, and automate paperwork in scripts and CI.
- [Agents](https://cloudraker.com/product/mcp.md): Connect CloudRaker to ChatGPT, Claude, Codex, or Grok and give your agent eleven real paperwork skills: parse, extract, classify, fill, redact, sign, and more.
- [Workspace](https://cloudraker.com/product/platform.md): See how CloudRaker turns AI output into governed operational work: product demos, feature cards, workflow controls, audit trails, and system actions.
## Industry solutions
- [Financial services](https://cloudraker.com/solutions/financial-services.md): Onboarding, credit approvals, exception handling, fund operations. Three-document reconciliation. The core banking credentials stay in the adapter. The reviewer keeps the decision.
- [Healthcare](https://cloudraker.com/solutions/healthcare.md): Clinical intake, referrals, prior authorization, patient correspondence. Field-level validation. The clinician keeps the clinical call. The EHR credentials never leave the adapter.
- [Insurance](https://cloudraker.com/solutions/insurance.md): Submissions, renewals, claims, servicing, certificates. The model proposes. Your rules decide. The policy admin system stays the source of truth.
- [Legal](https://cloudraker.com/solutions/legal.md): Matter intake, immigration files, document review, grounded knowledge retrieval. Every citation tied to firm records. The practice management credentials live in the adapter.
- [Manufacturing](https://cloudraker.com/solutions/manufacturing.md): Invoice intake, invoice matching, approval routing, supplier onboarding. The model proposes. Your rules decide. The ERP stays the source of truth.
- [Public sector](https://cloudraker.com/solutions/public-sector.md): Identity, case management, citizen services, municipal permitting. The model proposes. Your rules decide. The case management system stays the source of truth.
- [Real estate](https://cloudraker.com/solutions/real-estate.md): Lease extraction, CAM reconciliation, vendor invoice processing, vendor compliance activation, critical-date tracking. The model proposes. Your rules decide. The property management system stays the source of truth.
## Use cases
- [Medical scribe, built into your product](https://cloudraker.com/use-cases/medical-scribe.md): Audio in, structured clinical note out. CloudRaker transcribes, diarizes, extracts to your schema, cites every field back to the transcript, and stops for a human when the score is low.
- [Lease extraction across every document](https://cloudraker.com/use-cases/lease-extraction.md): Extract lease terms as cited JSON with unit across_documents. Several documents about one property fold into one record, deterministic rules pick the winner, and the review lists every disagreement.
## Models
- [decision-machine-1](https://cloudraker.com/models/decision-machine-1.md): decision-machine-1 is a decisions model from milliseconds.ai. Eight typed capabilities (yes/no, classify, rate, answer, extract, entities, verify) returning JSON with a probability attached in about a second, at $0.04 per million input tokens.
- [rakedoc-nano](https://cloudraker.com/models/rakedoc-nano.md): rakedoc-nano is CloudRaker's 1.2B-parameter vision model for document parsing. First among all vision-language models on ParseBench at 77.2, a bounding box on every element, open weights under AGPL-3.0, and $0.60 per 1,000 pages at enterprise volume.
- [rakeaudio-asr](https://cloudraker.com/models/rakeaudio-asr.md): rakeaudio-asr is the audio pipeline behind the CloudRaker Paperwork API. Transcription at 270x realtime, speaker separation at 150x, about 5% word error rate, and $5.00 per 1,000 minutes at every volume.
## Journal
- [Introducing decision-machine-1: a language model for decisions, not prose](https://cloudraker.com/journal/introducing-decision-machine-1-a-language-model-for-decisions-not-prose.md): decision-machine-1 is a language model that decides instead of writes. Eight typed decisions over text, one pass per call, a probability on every answer, $0.04 per million input tokens and 125M free every month. Now live at milliseconds.ai.
- [Give your agent hands: the paperwork CLI, skills, and MCP](https://cloudraker.com/journal/give-your-agent-hands-the-paperwork-cli-skills-and-mcp.md): Day 5 of launch week. The paperwork CLI, a skill pack for Claude Code and Codex, and an MCP server. Three layers that let a coding agent parse, extract, redact, fill and sign documents without ever loading them into context.
- [Introducing the Paperwork API: the whole platform, one surface](https://cloudraker.com/journal/introducing-the-paperwork-api-the-whole-platform-one-surface.md): The CloudRaker Paperwork API is six capabilities on one surface — extract, redact, compose, fill, redline, sign — with a citation beside every value, signed webhooks, SDKs in TypeScript and Python, and six runnable example apps on GitHub.
- [Introducing rakeaudio-asr: conversations are documents too](https://cloudraker.com/journal/introducing-rakeaudio-asr-conversations-are-documents-too.md): rakeaudio-asr is the audio pipeline behind the CloudRaker Paperwork API — transcription at 270x realtime, speaker separation at 150x, ~5% WER, and $5 per 1,000 minutes.
- [Introducing RakeSign: Free, Unlimited Document Signing](https://cloudraker.com/journal/introducing-rakesign-free-unlimited-document-signing.md): Announcing RakeSign: free, unlimited, legally binding document signing from CloudRaker. AATL-certified digital seals, a complete audit trail, SOC 2 Type 2 audited infrastructure, and a free sign API — because trust between humans and AI agents should be a commodity, not a cost.
- [CloudRaker Achieves SOC 2 Type 2 Compliance](https://cloudraker.com/journal/cloudraker-achieves-soc-2-type-2.md): CloudRaker has completed its SOC 2 Type 2 examination. An independent auditor verified our security controls operated effectively over time — here's what that means for your documents, and how to get the report on trust.cloudraker.com.
- [Introducing rakedoc-nano: a state-of-the-art open-weight VLM for document parsing](https://cloudraker.com/journal/introducing-rakedoc-nano-state-of-the-art-open-weight-vlm-for-document-parsing.md): rakedoc-nano is a state-of-the-art open-weight VLM for document parsing — 77.2 on ParseBench, first among all vision-language models, bounding boxes on every element, open weights under AGPL-3.0, and $0.60 per 1,000 pages at enterprise volume.
- [CloudRaker vs. DocuSign: The Signature Workflow on Canadian Infrastructure (2026)](https://cloudraker.com/journal/cloudraker-vs-docusign-signature-workflow-on-canadian-infrastructure-2026.md): A source-backed comparison of CloudRaker and DocuSign — Canadian data residency, redlining and the full agreement pipeline (parse, extract, redact, fill, sign), and how CloudRaker's credit model undercuts DocuSign's per-seat, per-envelope, per-module pricing.
- [CloudRaker vs. AWS Textract: A Full-Workflow Alternative (2026)](https://cloudraker.com/journal/cloudraker-vs-aws-textract-a-full-workflow-alternative-2026.md): Looking for an AWS Textract alternative? A source-backed comparison of CloudRaker and Amazon Textract: OCR, forms, tables, per-feature pricing, and the lifecycle stages (redaction, form-fill, e-signature, agent approval) CloudRaker adds.
- [CloudRaker vs. Datalab: OCR & Document Extraction API Comparison (2026)](https://cloudraker.com/journal/cloudraker-vs-datalab-ocr-document-extraction-api-comparison-2026.md): A source-backed comparison of CloudRaker and Datalab (Marker, Surya, Chandra): OCR and extraction accuracy, per-processor vs. one-pipeline pricing, and the lifecycle stages (redaction, e-signature, agent approval) CloudRaker adds.
- [CloudRaker vs. LlamaParse: Document Processing Comparison (2026)](https://cloudraker.com/journal/cloudraker-vs-llamaparse-document-processing-comparison-2026.md): A source-backed comparison of CloudRaker and LlamaParse — parsing and extraction, RAG vs. document-automation fit, the full lifecycle (redaction, form-fill, e-signature, agent approval), and pricing.
- [CloudRaker vs. Reducto: Document AI Platform Comparison (2026)](https://cloudraker.com/journal/cloudraker-vs-reducto-document-ai-platform-comparison-2026.md): A source-backed comparison of CloudRaker and Reducto as document AI APIs: parsing and extraction accuracy, the full document lifecycle (redaction, form-fill, e-signature, agent approval), pricing, and deployment.
- [CloudRaker vs. Google Document AI: A Full-Workflow Alternative (2026)](https://cloudraker.com/journal/cloudraker-vs-google-document-ai-a-full-workflow-alternative-2026.md): Looking for a Google Document AI alternative? A source-backed comparison of CloudRaker and Google Document AI: OCR, custom extraction, splitting/classification, per-processor pricing, and the lifecycle stages (redaction, form-fill, e-signature, agent approval) CloudRaker adds.
- [CloudRaker vs. CareWay: Build vs. Buy for the AI Medical Scribe](https://cloudraker.com/journal/cloudraker-vs-careway-build-vs-buy-for-ai-powered-clinical-documentation.md): CareWay is a finished AI medical scribe for clinicians. CloudRaker is the infrastructure product teams use to build and run scribe functionality inside their own platform — including a Canadian telehealth deployment that migrated stranded notes and expanded to patient consultations.
- [Giving Claude (or any agent) a paperwork tool](https://cloudraker.com/journal/giving-claude-or-any-agent-a-paperwork-tool.md): Two tools generated from the live OpenAPI spec instead of fifty hand-written ones. How CloudRaker's MCP server lets Claude, ChatGPT, or Codex drive a document from raw file to signature, and stop at the approval gate instead of guessing past it.
- [Inside the CloudRaker API: one pipeline, from raw PDF to signed document](https://cloudraker.com/journal/inside-the-cloudraker-api-one-pipeline-from-raw-pdf-to-signed-document.md): CloudRaker is the paperwork infrastructure. That's the tagline. This is what it means in code.
- [CloudRaker vs. stitching a document pipeline together yourself](https://cloudraker.com/journal/cloudraker-vs-stitching-a-document-pipeline-together-yourself.md): If your product touches contracts, claims, forms, or any document-heavy workflow, at some point someone on your team has had this conversation: "we just need to pull structured data out of this PDF and get it signed — how hard can that be?" It's never just that.
## Optional
- [Subprocessors](https://cloudraker.com/legal/subprocessors.md): The third parties that process customer data on CloudRaker’s behalf.
- [Trust center](https://trust.cloudraker.com): security posture, certifications, and subprocessor disclosures.
- [API documentation](https://docs.cloudraker.com/api/overview): endpoint reference for the document lifecycle API.
- [French site root](https://cloudraker.com/fr/): mirror of the above content in French, same `.md` convention applies.
