---

name: submit-directories description: Use when submitting a product to AI/startup directories - covers the full pipeline from collecting product info, analyzing directories, discovering forms, auto-submitting, handling captchas/OAuth/GitHub PRs, and tracking progress in checkpoint.md.

Submit Directories

Overview

Full lifecycle for submitting a product to 800+ AI tool directories: collect product info → analyze directories → build submission plan → discover forms → auto-submit → manual browser submissions → track progress.

Setup

Install dependencies

pip install -r requirements.txt
playwright install chromium

Configure credentials via environment variables

Never hardcode credentials in scripts. Set these before running any pipeline step:

# Required
export SUBMIT_PRODUCT_URL="https://yourproduct.com"
export SUBMIT_PRODUCT_NAME="Your Product Name"
export SUBMIT_TAGLINE="Your one-line tagline"
export SUBMIT_EMAIL="you@throwaway.com"        # use throwaway email
export SUBMIT_AUTHOR_NAME="Jane Doe"

# Recommended
export SUBMIT_AUTHOR_FIRST="Jane"
export SUBMIT_AUTHOR_LAST="Doe"
export SUBMIT_USERNAME="youruser"
export SUBMIT_PASSWORD="throwaway-pass"        # use throwaway password

# Optional
export SUBMIT_GITHUB_URL="https://github.com/you/repo"
export SUBMIT_TWITTER_URL="https://twitter.com/yourhandle"
export SUBMIT_KEYWORDS="ai,saas,marketing,automation"
export SUBMIT_LOGO="logo.png"                  # relative to script dir
export SUBMIT_SCREENSHOT="site-image.png"      # relative to script dir

Tip: save these to a local .env file (already in .gitignore) and load with:

set -a && source .env && set +a

Place assets

• logo.png — product logo (used for file upload fields)
• site-image.png — product screenshot (used for file upload fields)

Both should be in the same directory as the scripts.

Phase 0: Collect Product Info

Ask one at a time (never dump all at once):

#	Field	Question
1	Product URL	"What is your product/startup website URL?"
2	Product Name	"What is your product name?"
3	Tagline	"Give me a one-line tagline."
4	Description	"Write 2-3 sentences: what it does, who it's for, what makes it different."
5	Pricing	"What is your pricing model?" (Free / Freemium / Open Source / Paid)
6	Keywords	"List 5-7 category keywords."
7	Email	"What email for submissions?"
8	Name	"What name on submissions?" (get first + last separately too)
9	Username	"Preferred username for sites requiring registration?"
10	Password	"Throwaway password for sites requiring registration?" ⚠️ Use a throwaway — never a real password. Set via SUBMIT_PASSWORD env var; passwords are stripped from submission_plan.json before saving.
11	GitHub URL	(optional)
12	Twitter/X	(optional)

Then ask submission preferences:

• "Submit to Google-login directories? (You complete auth manually)"
• "Skip all paid directories, or flag for review?"
• "Fill forms on captcha sites and pause for you to solve, or skip?"

Assets: Tell user to place logo.png and site-image.png in project root.

Phase 1: Configure

1. Set all required environment variables (see Setup section above)
2. Generate 30 unique copy variations (title + description pairs):
   • Different angles: features, benefits, pricing, use case, comparison
   • Vary length: short punchy vs. detailed
   • Different keywords, never same opening
3. Store in submission_plan.json
4. Update checkpoint.md with product info (omit passwords)

Phase 2: Analyze & Classify

.venv/bin/python analyze_directories.py         # HTTP-level: auth, captcha, pricing, dead domains
.venv/bin/python cleanup_and_categorize.py      # Triage errors, build browser_check_list.json
.venv/bin/python browser_verify.py              # Playwright verification (10 concurrent workers)
.venv/bin/python browser_verify.py --recheck-unknown  # Deep recheck unknowns

Report auth type breakdown when done.

Phase 3: Build Submission Plan

Filter directories.json:

• site_status = active
• auth_type = none OR auth_type = email_password
• Include google_only / google_and_email only if user opted in

Create submission_plan.json entries with copy variation (rotate evenly), credentials, status: pending.

Phase 4: Discover Forms

.venv/bin/python discover_forms.py

Visits each submission URL, extracts form field metadata, updates submission_plan.json. Report: discovered / not found / timeout counts.

Phase 5: Auto-Submit

.venv/bin/python submit_directories.py

Report: submitted / no matching fields / timed out / needs manual attention.

Phase 6: Manual Browser Submissions

Use Playwright MCP (mcp__playwright namespace) for sites needing manual interaction.

Captcha sites

1. browser_navigate → submission URL
2. browser_snapshot → understand page
3. browser_fill_form / browser_type → fill all fields
4. Ask user to solve captcha
5. browser_click submit → verify confirmation

Google login sites

1. browser_navigate → login page
2. Click "Sign in with Google"
3. Ask user to complete Google auth
4. After login, proceed to submission form
5. Use browser_tabs to switch between OAuth popup and main tab

Rich text / complex forms

• TinyMCE/Quill: use browser_evaluate to set content directly
• Combobox/select: browser_click dropdown → click option
• Multi-step forms: browser_snapshot after each step for new fields
• Custom upload widgets: use browser_file_upload for logo/screenshot

Phase 7: GitHub PR Submissions

Requirements: gh CLI installed and authenticated (gh auth login). Set SUBMIT_GITHUB_URL if you want your repo linked in entries.

For awesome-list directories:

gh repo fork <owner>/<repo>
# Create branch, add product entry following repo format
gh pr create

Record PR URL in checkpoint.md.

Tracking

Update checkpoint.md after each phase with: counts by status, successful list, failed list with reasons, next steps.

Status Values

Status	Meaning
submitted	Confirmed submitted
skipped_paid	Requires payment
skipped_login_required	Requires account creation
captcha	Needs manual captcha solve
no_form_found	No submission form on page
no_fields_matched	Form exists, no fields matched
timeout / submit_timeout	Page/submit timed out
cloudflare_blocked	Cloudflare challenge
domain_parked	Dead domain

Rules

1. Never assume — unclear about product/preferences? Ask.
2. Never commit real passwords — use env vars only; passwords are stripped from submission_plan.json automatically before each save.
3. Report after every phase — never run pipeline silently.
4. Explain every skip — tell user why a directory was skipped.
5. Verify submissions — check for confirmation or error states.
6. Rate limit — max 5 concurrent submissions to avoid IP blocks.
7. Strip personal data before git push — search for email, name, password in all files.

Common Submission Blockers

Blocker	Frequency	Detection
Paid listing	~20%	Pricing page, Stripe links, "$" on submit page
reCAPTCHA / Turnstile	~10%	iframe[src*=recaptcha] or [data-turnstile]
Login required	~15%	Redirect to /login on submit URL
Newsletter-only forms	~10%	Looks like submit but is email signup
Domain parked/dead	~8%	No content, DNS failure
Cloudflare blocked	~3%	Challenge page, 403
Reciprocal link required	~5%	Old web directories requiring backlink

Best ROI Directory Types

1. AI tool directories with simple HTML forms
2. Startup directories using Google Forms
3. GitHub awesome-lists accepting PRs (free, high-quality backlinks)
4. NoCode/SaaS aggregators
5. General web directories with DA≥30

File Reference

File	Purpose
directories.json	Master database (827+ directories)
submission_plan.json	Targets with copy, fields, status (no credentials stored)
checkpoint.md	Progress tracking — source of truth
analyze_directories.py	HTTP-level analysis
cleanup_and_categorize.py	Triage + browser check list
browser_verify.py	Playwright browser verification
discover_forms.py	Form field discovery
submit_directories.py	Auto-submission engine