Claim Drafter User Guide

This guide provides detailed instructions for using all 11 modes of Claim Drafter. Follow the recommended workflow for best results, or jump to any mode as needed.

Recommended Workflow

Claim Drafter guides you through a structured patent claim development process. The recommended order is:

1. Determine Problem
2. Draft Claim 1
3. Draft Independent Claims
4. Extend to 15 Claims
5. Extend to as Many Claims as you can → Add numerous fallbacks in claim language. Useful as input for Web Drafter.
6. Add Reference Numbers
7. Generate Drawings → Create patent drawings from claims

Data transfers automatically between modes, so your outputs become inputs for the next step

Steps can be skipped

Make Drawings B/W → Independent mode that converts color images to patent-compliant B/W

Convert for Another Jurisdiction → Independent mode that rewrites an existing claim set to another patent office's formal requirements

Add Reference Numbers to Drawings → Independent mode with a manual editor for placing reference-numeral call-outs directly on a drawing image, plus an AI-assisted review pass

Edit Pictures → Independent mode for chat-based editing of any uploaded picture, which can then feed into the B/W conversion, reference-numbering, or figure export modes

Every main drafting mode accepts PDF, Word, or PowerPoint uploads, and every chat box and free-text field has a microphone button for voice dictation.

Mode 1: Determine Problem

Purpose

Analyze your invention disclosure to identify the objective technical problem solved, following the EPO problem-solution approach

Input

• Invention Description: Enter a technical description of your invention
• Document Upload (Optional): Upload a PDF, Word, or PowerPoint file with your invention description, drawings, or prior art. You can use text only, document only, or both together. The AI reads text and images from the file, and the PDF chat button lets you ask it questions directly.

Output

An objective problem statement and the AI's problem-solution reasoning behind it

Tips

• Edit by hand or use the chat feature to refine the problem statement

Mode 2: Draft Claim 1

Purpose

Generate your first independent claim based on the objective problem and invention description

Inputs

• Objective Problem: Auto-populated from Mode 1 (editable) or entered by you
• Description: Your invention description (synchronized across modes)
• Document Upload (Optional): Upload a PDF, Word, or PowerPoint file with additional details. You can use text only, document only, or both together.

Output

A first claim and the model's reasoning behind it

Tips

• Edit by hand or use the chat feature to refine claim 1
• Open a new window and run this mode with another model for a different proposal

Mode 3: Draft Independent Claims

Purpose

Generate additional independent claims covering different patentable aspects of your invention, including claims in other categories that may refer back to earlier claims where appropriate

Inputs

• Claim 1: Auto-populated from Mode 2 (editable) or entered by you
• Description: Your invention description
• Document Upload (Optional): Upload a PDF, Word, or PowerPoint file with invention details. You can use text only, document only, or both together.

Output

Multiple independent claims covering all aspects the model considers adequate, as well as the model's reasoning behind its choice. Claims in different categories may refer back to earlier claims where that is the cleaner drafting choice:

Tips

• Edit by hand or use the chat feature to refine these claims
• Open a new window and run this mode with another model for a different proposal

Mode 4: Extend to 15 Claims

Purpose

Build a complete claim set by adding dependent claims to your claims, while preserving valid cross-category references

Inputs

• Claims: Auto-populated from Mode 2 or 3 (editable) or entered by you
• Description: Your invention description
• Document Upload (Optional): Upload a PDF, Word, or PowerPoint file with invention details. You can use text only, document only, or both together.

Output

A 15-claim set with:

• All your input claims preserved
• Dependent claims properly narrowing each independent claim
• Valid cross-category references maintained
• Sequential numbering (1-15)
• Proper dependency chains

Tips

• Edit by hand or use the chat feature to refine these claims

Mode 5: Extend to as Many Claims

Purpose

Generating additional dependent claims from your invention description.

Inputs

• Claim Set: Auto-populated from Mode 2, 3 or 4 (editable) or entered by you
• Description: Your invention description
• Document Upload (Optional): Upload a PDF, Word, or PowerPoint file for additional context. You can use text only, document only, or both together.

Output

An extended claim set with:

• All previous claims preserved
• Additional dependent claims covering fallback positions for prosecution

Tips

• Use this extended claim set as input in Web drafter to generate fallbacks in the description
• These claims can be added during prosecution if needed
• Edit by hand or use the chat feature to refine these claims

Mode 6: Add Reference Numbers

Purpose

Insert parenthetical reference numbers into your claims to coordinate with patent drawings.

Input

• Claim Set: Auto-populated with the most comprehensive claim set available (extended > 15 claims > independent > claim 1)
• Document Upload (Optional): Upload a PDF, Word, or PowerPoint file with drawings. The AI can extract reference numbers from images. You can use text only, document only, or both.

Output

Claims with reference numbers in parentheses, e.g.:

• "a housing (10) comprising..."
• "a motor (20) connected to the housing (10)..."
• "a sensor (30) positioned within..."

Tips

• If you desire specific numbers, give instructions to this respect at the end of your input or in the chat feature after the generation
• Edit by hand or use the chat feature to refine the results

Mode 7: Generate Drawings

Purpose

Create patent-style drawings directly from your claims with reference numbers, as SVG or PNG.

Input

Claim Set: Claims with reference numbers from Mode 6

Output

Patent drawings with:

• Schematic representations of claim elements
• Reference numbers matching the claims
• Multiple figures for different views/embodiments

Features

Format Choice	Generate as SVG, or as PNG using a separate AI/commercial image model
Prev/Next Navigation	Browse multiple generated figures
Edit Mode	Click Edit to drag-and-drop elements, resize shapes, double-click text to modify (SVG output)
Download	Export the current drawing as an SVG file (editable in PowerPoint) or a PNG file
Chat Refinement	Request specific modifications via the chat panel, for either format
Undo	Revert the last chat-requested change

Tips

• Generation may take a moment for complex inventions
• SVG output gives you the interactive Edit mode; PNG output goes through chat refinement instead
• Use Edit mode for interactive adjustments
• SVG files can be further edited in PowerPoint after 'ungrouping'
• Use a reasonably low number of claims (1-5) as input
• Chose well the claims you wish to be drawn

Independent Mode: Make Drawings B/W

Purpose

Convert color images to patent-compliant black and white line drawings. This is an independent mode that does not require completing the main workflow.

Inputs

• Drawing Files: Upload one or more color images, or a PDF, Word, or PowerPoint file to extract pictures from (multiple file selection supported)
• Relief Option: Enable for 3D depth with hatching and stippling
• Remove Wording (default on): Cleanly strips existing text and reference numbers from the drawing so you can add your own in Add Reference Numbers to Drawings
• Instructions (Optional): Free-text guidance for this run, e.g. what to keep or what to remove
• Model Selection: Use the visible B/W Conversion Model dropdown. EU-hosted open-source is one of the available options, with additional models available depending on your account.

Output

Patent-compliant black and white line drawings suitable for filing.

Features

3D Relief	Adds hatching and stippling for depth perception
Model Choice	EU-hosted open-source is one of the available paths, with frontier or specialist alternatives available depending on your account.
Remove Wording	Cleanly strips existing text and reference numbers so you can add your own
Seed Options	Generate 3 style variations to choose from for a single drawing
Generate 3 Versions	Generate 3 full versions of the batch to compare side by side
Batch Processing	Convert multiple images with consistent styling
Per-Figure Regenerate	Re-run a single figure without redoing the whole batch
Show Original	Compare a converted figure against its original at any time
Batch Chat Edit	Select several figures and apply one chat instruction to all of them at once
Undo	Revert the last change, including a batch edit

Tips

• Check carefully the convertion for errors
• Chat with the results to modify the drawings or add reference numbers
• Try entering a claim with reference numbers in the chat and ask the model to add those to the drawing
• If the input has hand-written reference numbers, they will be converted into printed ones
• Relief works best for converting edgeless 3D features; disable for most uses
• Turn on Remove Wording when you plan to add your own reference numbers afterward in Add Reference Numbers to Drawings
• Use Instructions to keep or remove specific labels the model might otherwise handle inconsistently
• Tick several figures and use batch chat edit to apply the same fix everywhere at once, instead of repeating it figure by figure
• Select figures and use "Add reference numbers" to send them on to the manual reference-numbering editor, or "Add to figure set" to include them in your export

Independent Mode: Convert for Another Jurisdiction

Purpose

Rewrite an existing claim set to the formal requirements of another patent office. This is an independent mode that does not require completing the main workflow: select Convert for another jurisdiction in the mode dropdown.

Inputs

• Claim Set: Paste the claims to convert (any language)
• Target Jurisdiction: USA, EPO, China, Japan, South Korea, Canada, Brazil, India, Australia, Mexico, Singapore, or Indonesia
• Output Language: Same as input, English, or the language required by the target jurisdiction

Output

The converted claim set, together with an info panel showing:

• Rules applied: the target jurisdiction's claim-format requirements used for the conversion (dependency structure, multiple dependent claims, two-part form where required, and so on)
• Claim fees: an excess-claim fee summary for your claim count (e.g. fees above 20 claims in the US, from claim 16 at the EPO, above 10 in China), with the date the fee rules were last checked

Tips

• Review the converted set carefully — jurisdiction rules can require restructuring dependencies, which renumbers claims
• Substantive claims are never silently deleted to get below a fee threshold; consolidate manually if you want a cheaper claim set
• Fee amounts change regularly — check the "as of" date in the fee summary and verify current official fees before filing
• Use the chat panel to refine the converted claims, as in every other mode

Independent Mode: Add Reference Numbers to Drawings

Purpose

Place reference-numeral call-outs directly onto a drawing image, with a manual editor and an AI-assisted review pass. This is separate from Mode 6, which inserts reference numbers into claim text—this mode places them onto the figure itself. It is an independent mode: select Add reference numbers to drawings in the mode dropdown, or send figures to it directly from Make Drawings B/W or Edit Pictures.

Inputs

• Drawing Files: Upload one or more black-and-white drawings, or send figures in from Make Drawings B/W or Edit Pictures
• Claim Set with Reference Numbers (Optional): Paste the claims so the editor knows which numbers to place

Output

Annotated drawings, downloadable as PNG or SVG, with:

• Reference numbers placed on or beside each feature
• Leader lines connecting numbers to features where needed
• Multiple figures browsable with Prev/Next navigation

Features

Manual Placement	Drag numbers and their leader-line endpoints anywhere on the figure
Leader Styles	Straight, curved, or double-curved leaders, with an option to flip the curve direction
Arrowhead Options	None, at the feature end, at the number end, or both
Underline / On Feature	Print a bare underlined number directly on a feature instead of drawing a leader to it
Curly-Brace Grouping	Group several features under a single number
Multi-Select	Shift-click to select several numbers, then move, delete, duplicate, or copy/paste them as a group
Size & Thickness	Adjust reference-number size and leader-line thickness, per figure or across the whole set
AI Review	Runs a check on the placement and applies smart, uncluttered corrections that automatically adapt to your figure; can also run automatically after each figure is planned
Undo	Revert the last change in the editor

Tips

• Let the editor place numbers automatically first, then use the manual tools to adjust only what needs it
• Run "Review with AI" (or tick "Auto-review after planning") after manual edits to catch crowding or crossed leaders
• Use On Feature placement for large, otherwise-blank features instead of adding another leader
• Multi-select is useful for moving or duplicating a group of related numbers together
• Once your figures are annotated, add them to the figure set below for export

Building Your Figure Set

Figures collected across the drawing modes appear in a shared figure-set tray for arranging and exporting:

• Reorder: Drag figures into the order you want
• Labels: Automatic figure labels (e.g. "Fig. 3A"), or set your own and they stick through reordering
• Export as Word: A4 pages, one figure per page, styled captions and page header, with an option to place a figure on a sideways page
• Export as ZIP: A .zip archive of the individual image files

Independent Mode: Edit Pictures

Purpose

Chat-based editing for any uploaded picture, not just claim drawings. This is an independent mode: select Edit Pictures in the mode dropdown.

Inputs

• Picture Files: Upload images directly, or a PDF, Word, or PowerPoint file—each page lands in a staging tray where you can select, crop, or split it before loading it into the editor

Output

Edited pictures, ready to send onward to other modes.

Features

Staging Tray	Thumbnails with select, crop, and split tools before loading pictures into the editor
Chat Editing	Edit a single picture via chat
Batch Chat Edit	Select several pictures and apply one chat instruction to all of them at once
Undo	Revert the last change, including a batch edit
Send to Convert to B/W	Send selected pictures on for black-and-white conversion
Send to Add Reference Numbers	Send selected pictures on to the reference-number annotation editor
Add to Figure Set	Send selected pictures straight into your exportable figure set, without conversion

Tips

• Use Edit Pictures to prep photos, screenshots, or scanned figures before sending them into the B/W or reference-numbering modes
• Select several pictures and batch chat edit to apply the same change to all of them, instead of repeating it one by one
• Pictures added to the figure set here skip B/W conversion—use this for figures that are already patent-ready as-is

Settings & Options

Model Selection

Click "Advanced Options" to reveal the model dropdown for the main drafting modes and Generate Drawings:

• EU-hosted open-source – Qwen/Mistral path on EU infrastructure
• Gemini – Frontier option, where enabled for your account
• GPT – Frontier option, where enabled for your account
• Claude – Frontier option, where enabled for your account

Note: Generate Drawings uses this same Advanced Options model dropdown. Convert Drawings to B/W has its own visible B/W Conversion Model dropdown instead.

Expert Drafting Options

Advanced Options also exposes a few extra controls depending on the mode and model in use:

• Drafting Profile: A Fast / Balanced / Thorough slider that trades speed for how much time the model spends on a given generation
• Enhanced-Accuracy Mode: An optional higher-accuracy mode for high-stakes claim 1 drafting, for cases worth the extra processing time
• Two-Part Claim Format: A checkbox to toggle two-part ("characterised in that") claim format on or off for Claim 1
• Chemistry Mode: A checkbox that switches Claim 1 drafting to a mode tuned for chemistry claims

Language

Specify your desired output language (e.g., "German", "French"). Leave empty for English.

Document Upload & PDF Chat

Every main drafting mode accepts a PDF, Word, or PowerPoint file alongside or instead of typed text. Once a document is uploaded, a "PDF chat" button lets you ask questions about it directly instead of only using it as drafting context.

Voice Dictation

Every chat box and free-text field carries a microphone button. Click to record, click again to stop; the transcript (produced by an EU-hosted speech-to-text service) is inserted at the cursor without being sent automatically, so you can correct it first.

Saved Projects

Claim Drafter autosaves your work continuously as you go—there is no manual save step. From the project bar you can:

• New: Start a fresh project
• Rename: Give the current project a name you'll recognize later
• Duplicate: Copy a project to branch off a variant
• Delete: Remove a project you no longer need

Reopening a saved project restores your claims, drawings, annotations, and figure set.

Credits & Time Remaining

The header displays your current credit balance and time remaining:

• Credits: Your available balance for using the application.
• Time Remaining: Once you use the tool, 0.4 credits will be deducted. The tool can then be used without limitations for 24 hours. This countdown shows how much time is left before the used end of this 24h period.

Live Progress

Longer operations—claim generation, drawing conversion, reference-number placement—show a live progress indicator with the current phase and step, so you can see the tool is working rather than wondering if it has stalled.

Chat Refinement

After generating any output, use the chat panel at the bottom to request modifications to the output. For text outputs, AI-proposed edits appear as a track-changes-style review: accept, reject, or reset each change individually, or accept or reject them all, before applying them to your claims.

Tips for Best Results

Input Quality

• Clearer invention descriptions yield better claims
• There is no limit on the input length
• Mention prior art limitations you want to overcome
• PDF, Word, and PowerPoint uploads can include text, drawings, or both—the AI reads everything
• Use text only, document only, or combine both for maximum context
• Dictate instead of typing in any chat box or text field with a microphone button

Workflow Efficiency

• Review and refine at each step before proceeding
• Use chat to make adjustments rather than restarting
• Review chat-proposed edits in the track-changes panel before accepting them
• Open Claim Drafter in a new window and use advanced options to try other models on the same input
• Data transfers automatically—no copy-paste needed
• Save projects as you go so you can branch variants with Duplicate instead of redoing work

Drawing Quality

• Use Edit mode for fine adjustments on SVG drawings
• Download as SVG for further editing in other tools, or PNG for a photorealistic figure
• Turn on Remove Wording before adding your own reference numbers, so old labels don't linger
• Let the reference-number AI review pass clean up placement, then adjust the remaining figures by hand
• Build your figure set as you go, and export it as a Word document or ZIP archive when your drawings are final