AI-native automation
Automate work across
GitHub and 
PDFwith Caddi AI Agents.
Connect GitHub and PDF: Document extract in PDF when new issue in GitHub. Teach Caddi the job once, the way you'd train someone new, and it runs end to end. No workflow builder.
Discover, automate and govern the work between GitHub and PDF
Caddi AI Agents find the repetitive handoffs between GitHub and PDF, learn them the way a new hire would, and run them for you, with every run on the record. No workflow builder, no code.
Discover repetitive work
across all your tools.
Train an AI agent to do the work
over a video call.
Search mailMatters / Whitmore HoldingsGovern every AI agent,
every decision.
Discover repetitive work
across all your tools.
Train an AI agent to do the work
over a video call.
Search mailMatters / Whitmore HoldingsGovern every AI agent,
every decision.
Discover repetitive work
across all your tools.
Train an AI agent to do the work
over a video call.
Search mailMatters / Whitmore HoldingsGovern every AI agent,
every decision.
What Caddi is and how it works
An AI teammate that runs your back-office loops.
- Doesn't break. Caddi reads intent, so when fields move or UIs change, your loop keeps running.
- Taught like a new hire. Walk Caddi through the work once. Tweak it later by chat, with no workflow builder to re-architect.
SOC 2 attested · Human-in-the-loop · Full audit trail · 150+ tools across legal, finance, and operations
Practical ways to use GitHub and PDF together
- 01
Document extract in PDF when new issue in GitHub.
Caddi watches GitHub for new issue and document extract in PDF. No copy-paste, no missed records.
- 02
Create issue in GitHub when classify in PDF.
Caddi watches PDF for classify and create issue in GitHub so the two systems stay in lockstep.
- 03
Extract in PDF from GitHub events.
When pull request merged happens in GitHub, Caddi extract in PDF with the right context attached.
Actions Caddi can take across GitHub and PDF
- GitHub
New issue
Triggers when an issue is opened in a repository.
- GitHub
Pull request merged
Triggers when a pull request is merged.
- GitHub
New release
Triggers when a release is published.
- GitHub
Create issue
Open an issue with labels and assignees.
- GitHub
Comment on issue
Add a comment to an issue or pull request.
- GitHub
Update labels
Add or remove labels on an issue.
- PDF
Classify
Classifies text content using AI with customizable classification types (language, tone, sentiment, topic, category), accepting content directly or automatically downloading from S3.
- PDF
Summarize
Summarizes text using AI with customizable style and length parameters, accepting content directly or from S3.
- PDF
Get form fields
Discovers all form fields in a PDF including field types, options, and validation rules.
- PDF
Document extract
Extracts text or structured fields from PDF documents and images stored in S3 Use mode "full_text" to extract all document text, or mode "schema" with a JSON Schema to extract specific named values…
- PDF
Extract
Extracts specified information from text using AI, accepting content directly or automatically downloading from S3.
- PDF
Extract data from PDF
Extracts specific field values from PDF documents using AWS Textract for OCR and LLM for intelligent field mapping, with automatic batching strategy for large schemas.
- PDF
Extract fields with OCR
Extracts specified field values from images (JPEG, PNG) and PDFs using AI-powered OCR with confidence scoring and validation, automatically downloading content from S3 (supports both full S3 keys a…
- PDF
Extract multiple fields
Extracts values for specified fields from text using AI, accepting content directly or from S3.
- PDF
Write
Generates written content using AI with customizable style, format, tone, length, language, purpose, and audience parameters, accepting content directly or automatically downloading from S3.
- PDF
Extract pages
Extracts one or more pages from a PDF stored in S3 and saves the result as a new PDF file in S3 Specify start_page and end_page (1-indexed, inclusive) to select the page range.
- PDF
Extract text
Extracts all text from PDFs with support for page ranges and maintains text structure.
- PDF
Fill form
Fills PDF form fields and automatically saves the completed PDF to S3 with optional flattening.
- PDF
Image to PDF
Downloads images from S3, converts them to a multi-page PDF with configurable page sizing, and uploads the result to S3.
- PDF
Merge documents
Downloads multiple PDFs from S3, merges them with optional page range selection, and uploads the merged PDF to S3 Set flatten_form_fields to true when merging PDFs with filled forms or signatures t…
- PDF
Text to PDF
Creates a PDF from text content with configurable formatting options (page size, margins, font size, bold text), and uploads the result to S3.
Common questions
How does Caddi connect GitHub and PDF?
GitHub and PDF just run together. You teach Caddi the way you'd teach a new hire: walk it through how you use them today, with no workflow builder to wire up. Caddi turns that walkthrough into a verified loop and runs it against GitHub and PDF end-to-end.
Do I need engineering help?
No. Whoever does the work today walks Caddi through it on a call, the same way they'd train a new hire. Caddi builds the loop. No code, no IT ticket.
Is my data safe?
Yes. Caddi is SOC 2 attested. The AI watches and builds the automation, and once the loop ships it runs as deterministic code, repeatable and logged, with any model step scoped and reviewable.
Can Caddi connect GitHub and PDF to other tools too?
Yes. Most Caddi loops span 3–6 tools. Once the first GitHub + PDF loop is live, Caddi suggests other tools and other workflows to fold in.
How fast can it go live?
Typical first loops ship within a week of that first walkthrough. The Caddi team reviews every loop before it runs in production.
Ready to automate GitHub and PDF?
Drop your work email and we'll show you Caddi running end-to-end against GitHub, PDF, and the rest of your stack.
Or skip the demo: sign up free and build it yourself.