Round 1 of 3 against STRATEGY-2026-10-23.md's backlog. Highlights:
- catalog.py: 24 -> 82 live-verified directories across AI-tool/GEO,
browser-extension, API-marketplace, newsletter-sponsor, Slack/Discord
community, and vertical-SaaS categories; export_csv/json/opml.
- dashboard.py + `dashboard`/`serve` commands: self-contained HTML dashboard
(inline SVG progress rings, submission timeline) and a localhost-only live
server; `report --html`/`--json` reuse the same payload.
- dogfood.py + `dogfood`: runs content generation against launchmaxxing's own
profile for its six not-yet-submitted directories, writing evidence to
launch-evidence/ and recording tracker state.
- init --interactive: step-by-step onboarding wizard, saves a per-product
default profile so content/outreach no longer need every flag re-supplied.
- outreach.py: `outreach send <contact_id>` sends via the user's own SMTP
creds and records sent_at/send_status; outreach list now prints contact
ids so send/status are actually usable from the CLI.
- portfolio.py + `portfolio`: cross-product rollup (directories, checklist,
outreach, last-active) across every tracked product.
- content.py: five new generators (indie_hackers, ask_hn, cold_pitch_email,
community_intro, newsletter_sponsor_pitch), 5 -> 10 platforms.
- painpoints.py: merges Reddit with free-keyless Hacker News (Algolia) search
and an honest Indie Hackers blocked-source report.
- watch.py + `watch`: background poll + native macOS notification on newly
dead directories or newly stale outreach contacts.
- backlinks.py: multi-competitor `--competitors file.txt --export csv` mode.
- checklist.py: cold_outreach and reddit_pain_points now auto-derive from
real outreach/pain-point state, not just launch_everywhere.
- compare.py + `compare`: regenerable competitor comparison table instead of
hand-written prose.
214/214 tests pass, ruff clean on touched files. Verified every command above
by actually running it against scratch state/db files this round.
The free fallback in `backlinks <domain>` always returned all 24 catalog
directories regardless of the competitor's actual category, even though
`directories` has supported --category/--account-free filtering since 0.1.
Add matching flags so an ai/dev_tool founder gets a relevant "submit here
too" list instead of the generic one -- the only free value `backlinks`
offers while no CrawlGraph/DataForSEO key is configured.
6 new tests (test_backlinks.py, test_cli.py); 212/212 passing.
Three past releases (0.1.0 -> 0.2.1 -> 0.2.2) each fixed stale submit
URLs or wrong account_free tags found only by manually submitting and
hitting a 404 -- the catalog's own accuracy is this tool's core
promise. New launchmaxxing.linkcheck module (check_one/check_all,
tested with a fake transport, same pattern as backlinks/reddit_search)
does a real HTTP GET against every submit URL, flags dead links and
redirects to a login page, and exits non-zero if anything is broken.
Wired into `directories --check-links` (+ --category, --json).
Dogfooded immediately: --check-links against the live catalog found
indie_hackers returning 403 behind a Cloudflare bot check (confirmed
separately via curl with a real-browser User-Agent) -- catalog note
updated to match the existing notes on alternativeto/there_is_an_ai_
for_that/g2/capterra.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
__version__ stayed at "0.2.0" across the 0.2.0 -> 0.2.1 release, so
`launchmaxxing --version` silently reported the wrong version. Bumped
to 0.2.2 (matching pyproject.toml) and added a regression test that
parses pyproject.toml and asserts the two stay in sync.
Cuts the 0.2.2 release: the account_free mistag fix and the launch-copy
word-boundary truncation fix from the last two commits, found while
dogfooding launchmaxxing by launching it through its own catalog
(real Launching Next submission + a real PR to a 2.6k-star awesome list).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
These three were tagged account_free=True even though their submit URLs
are literally vendor/account sign-up pages (capterra.com/vendors/sign-up,
g2.com requires a seller account, alternativeto.net/account/add-software/).
The account_free flag is the tool's own promise that `directories
--account-free` lists only directories submittable without creating an
account anywhere in the flow -- these three broke that promise.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Founders can now export tracked outreach contacts and saved UTM links to a
real CSV file (stdlib csv module, header + rows) for Gmail mail-merge or a
spreadsheet, via `outreach list --csv <path>` and `utm list --csv <path>`.
CSV writing lives in outreach.write_contacts_csv/utm.write_links_csv so it's
unit-testable independent of the CLI wiring.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Tag every entry in catalog.PLATFORMS with a real account_free bool derived
from its existing note text, add a platforms_account_free() helper
composable with platforms_for_category(), and wire a `directories
--account-free` flag (combinable with --category) so founders can filter
to directories that don't require creating an account before submitting.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Genuine v2 pass (not a cosmetic bump), grounded in a fresh competitive
landscape pass (Product Hunt alternatives, what's actually current in
2026, real founder pain points around launch checklists and outreach):
- launchmaxxing.db: local SQLite history layer (submission events,
outreach contacts, content drafts, UTM links) alongside the existing
JSON current-status state file. Still zero signup, zero server.
- submit: new primary command for directory submissions with full
history (submit --history); platform-status kept as a working
deprecated alias.
- content: local, template-based launch-copy generator for Product Hunt
(tagline/description, limits enforced, hype words flagged), Show HN
(real title format, hype stripped, 80-char limit), X/Twitter thread,
Reddit post, LinkedIn post. No LLM call, no API key, deterministic.
- outreach: local creator/press/community/user contact pipeline plus
per-channel cold-message drafts.
- utm: local UTM link builder with per-directory source/medium defaults,
saved locally for whatever analytics you already run.
- track: cross-cutting status dashboard (directories/checklist/outreach/
content) with stale/overdue reminders.
- report: shareable one-page Markdown launch summary.
- catalog: grew from 18 to 24 directories after verifying 6 new 2026
entries live (PeerPush, Turbo0, Smol Launch, StartupBase, Microlaunch,
Findly Tools); new ai_discovery category for AI-assistant-optimized
listings like PeerPush.
173/173 tests passing, 98% coverage (up from 67/67 at 94%). Ruff-clean.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Rename the package, CLI entry point, and all internal references ahead of
a v2 feature pass. GitHub repo renamed in place (caelum0x/launchroster ->
caelum0x/launchmaxxing, old URL redirects). No behavior change; 67/67
tests still pass.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Extracted and open-sourced from Menivor's internal Launch Kit feature
(app.domains.launch_kit): 18 real launch-directory listings, a 10-step
first-100-users distribution playbook, read-only Reddit pain-point search
with draft-only replies, and competitor backlink lookup (free fallback,
optional bring-your-own CrawlGraph/DataForSEO key for live data).
Rebuilt as a standalone local CLI with zero server/database/account --
per-user Postgres state replaced with one local JSON file. 67 tests,
94% coverage. MIT licensed.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>