Large repos do not fit cleanly into chat windows
Audits and reviews often span source files, tests, configs, docs, interfaces, and generated artifacts. Pulling the right subset by hand is slow and easy to get wrong.
The context problem is workflow debt.
LLM review gets weaker when the input is improvised. PastePrompt focuses on repeatable context assembly before anything is pasted into an external model.
Manual copy/paste loses the review trail
Once context is pasted into an LLM tool, it is hard to remember which files, diffs, prompts, and assumptions were included.
Bad selections waste tokens and attention
Wrong files crowd out the code that matters. Missing dependencies force follow-up prompts and make results harder to compare.
Accidental secrets are a real workflow risk
Context bundles should be reviewed before they leave the machine. Ignore rules and local scanning reduce accidental inclusion.
A focused flow from local repo to reviewable bundle.
The checklist maps the full context-building path. The walkthroughs below are actual app recordings from a local demo repository.
Select a local repo
Open the codebase from disk and keep repository contents local.
repo: ./client-appSearch and select files
Filter large trees and select only the source, tests, or docs needed for the review.
select: src/core/**/*Add local imports
Use local dependency suggestions, then include tests, docs, or configs that explain the selected code path.
imports: local helpersReview token budget
Check token estimates by file, folder, and selected context before export.
tokens: 42k / 64kAdd Git diff context
Include changed files and metadata for PR review or patch analysis.
diff: working treeApply review prompt template
Use repeatable prompts for security review, regression review, architecture mapping, or diff review.
template: reviewRun secret scan
When enabled, review likely secrets before copying or exporting the bundle.
scanner: review passedCopy or export bundle
Produce XML or Markdown for ChatGPT, Claude, Codex, Cursor, Gemini, or another LLM tool.
export: context.mdSelection
Bundle preview
Secret scanner
Watch PastePrompt in the real app.
These short videos were recorded from the installed macOS app. They use a local demo repository and do not show customer code.
Actual app recording8 sec
Select files and watch the token budget
Search a local repository, select matching files, and confirm the selected token estimate before building context.
Actual app recording16 sec
Preview, copy, and preserve history
Generate an XML-like bundle, pass the local preflight check, copy context, and see the saved prompt-history metadata.
Actual app recording12 sec
Review real Git diffs
Switch to Git diff mode, load unstaged changes, select changed files, and include diff context in the generated bundle.
Actual app recording10 sec
Catch secrets before copying
Trigger the local scanner with a synthetic demo key and review redaction, exclusion, cancel, and copy-anyway choices.
The scanner walkthrough uses a synthetic demo key so the warning flow can be shown safely.
Core features for large-codebase context work.
The MVP is built around selecting the right material, checking it locally, and exporting it in a format that external LLM tools can use.
Local macOS app
Desktop-first workflow for selecting and packaging local repositories.
Fast file tree
Navigate large codebases without treating the whole repo as one blob.
Search and filters
Find source files, tests, configs, docs, and review targets quickly.
Token counts
Estimate size by file, folder, and selected bundle.
.gitignore and .pastepromptignore
Respect standard ignores and add context-specific exclusions.
Syntax-highlighted preview
Inspect source before it enters the generated context.
XML/Markdown bundles
Generate structured XML or reviewable Markdown context.
Prompt templates
Reuse audit prompts for recurring review modes.
Workspaces
Keep repeatable selections and settings for active codebases.
Secret scanner
Run a local safety check before copy/export when enabled.
Git metadata
Include branch, commit, and repository state when it matters.
Git diff mode
Prepare context around changes under review.
Local dependency resolver
Suggest nearby source imports for selected files.
Prompt history
Track generated prompt metadata without storing raw source by default.
Markdown export
Save context bundles as `.md` artifacts.
Open in VS Code/Cursor
Jump from selected context back to the editor.
Why local-first matters.
PastePrompt is designed for sensitive repositories and professional review workflows. It does not require uploading source code to a PastePrompt backend for V1.
Read the security modelLocal selectionChoose files from a repository on your Mac.
Ignore-aware by defaultRespect `.gitignore` and `.pastepromptignore` before context generation.
Scanner before copy/exportWhen enabled, review likely secrets before generated context leaves the app.
User-controlled sharingContext leaves when you copy, export, paste, attach, or share it.
Built for audit workflows.
PastePrompt is not a vulnerability scanner. It helps auditors and reviewers prepare better input for the analysis they already perform.
Security reviews
Package source files, tests, configs, interfaces, docs, and focused review prompts for security analysis.
Security research
Prepare repeatable snapshots for attack-surface mapping, dependency review, and hypothesis-driven investigation.
Code review and PR analysis
Use Git diff mode with surrounding source files so reviewers can ask targeted questions about changed behavior.
Consultant handoff
Export Markdown bundles that document what was reviewed, which files were included, and what prompt shaped the review.
How it works.
Build context like an artifact: select, inspect, check, generate, and reuse.
- SelectOpen a local repo and choose files, folders, related context, and diffs.
- InspectPreview source and keep token budget visible before export.
- CheckRun local secret scanning and adjust selections when needed.
- GenerateCopy XML or export Markdown for your chosen LLM workflow.
Supported workflows.
PastePrompt is useful anywhere selected code context needs to be inspected, reproduced, and moved into external LLM tools.
BYO LLM tools
Copy or export context for ChatGPT, Claude, Gemini, Codex, Cursor, and other tools that accept pasted or attached code context.
Large repository triage
Start with a broad tree, narrow by search and filters, then export only the files that support the question.
Repeatable audit prompts
Keep templates for recurring review patterns instead of rewriting prompt framing for every repo.
Reviewable artifacts
Use Markdown exports when the context bundle itself needs to be stored, reviewed, or shared internally.
How PastePrompt differs from generic context tools.
The difference is workflow shape: local selection, audit-oriented metadata, scanner gating, and reproducible bundles.
ApproachTypical tradeoffPastePrompt focus
Manual copy/pasteFast for one file, but fragile across many files, diffs, prompts, and repeat reviews.Keeps selection, token planning, templates, scanning, and export in one local workflow.
Generic repo packersOften useful for dumping files, but less focused on review prompts, diffs, scanner gating, and reproducible snapshots.Built around review-specific context bundles and before-export checks.
Editor-only extensionsConvenient inside one editor, but can be tied to editor state and may not produce portable artifacts.Creates standalone XML or Markdown bundles while still opening selected files in VS Code or Cursor.
AI coding agentsGood for interactive work, but not always suited to documenting exactly what context was packaged for a review.Prepares reproducible inputs for the LLM or agent workflow you already choose.
Launch pricing.
Founder and Pro purchases are handled by email during launch so terms, invoices, and license delivery can be confirmed manually.
Founder Lifetime
€99
Founder access for customers who want a long-term personal license and a direct support path during launch.
Pro Annual
€149/year
For auditors, reviewers, and consultants who use repeatable context-building workflows throughout the year.
Pro Monthly
€19/month
Monthly access for shorter engagements, trials, and teams evaluating whether PastePrompt fits their workflow.
FAQ.
Short answers for launch users. Keep the full docs authoritative for install, license, and release details.
Does PastePrompt upload my source code?
PastePrompt is designed as a local-first macOS app. Source code leaves your machine when you copy, export, paste, attach, or otherwise share a generated bundle.
Does it replace an auditor?
No. PastePrompt prepares context. It does not find vulnerabilities by itself and does not guarantee LLM output.
Which LLMs does it work with?
It supports BYO LLM workflows by copying or exporting context for tools such as ChatGPT, Claude, Gemini, Codex, Cursor, and others.
Is it tied to one programming language?
No. PastePrompt works at the repository and file-selection layer, so it can prepare context for many language stacks and review workflows.
Is there a Windows version?
The initial product focus is macOS. Do not assume a Windows release until it is announced in the release notes.
Build the context first. Then ask the model.
Use PastePrompt to package code, diffs, related files, prompts, and checks into bundles you can inspect before sending to your LLM workflow.