LaunchRepo vs backlink-pilot: free script or curated kit

By Aleksandar Vuk JovanovicHow we research and update

Checked on October 1, 2026 — prices and features from the public website of backlink-pilot; verify before buying.

Short answer: backlink-pilot is free, MIT-licensed code that fills directory forms from one config file, and it also ships a module that posts blog comments in batches. LaunchRepo is a repository you buy once: 299 curated playbooks your own agent works through, with rules that keep you within each platform’s terms and a tracker that separates submitted from live. Pick backlink-pilot to tinker for free; pick LaunchRepo when you want a launch process you can trust with your product’s name.

backlink-pilot is a free, open-source toolkit on GitHub that submits a product to directories from one config file. It drives your own Chrome through bb-browser, counts 259 target sites in its README, and can be set up by Claude Code. LaunchRepo also works in your own browser with your own agent, but it is built the other way round: curated playbooks first, automation second. We read backlink-pilot’s canonical repository, its README, target list and source code, without running it, and compared both on what the files actually contain.

Five questions that decide it

Answer each question for the product you are launching next. Where most of your answers land is your pick.

Question	Pick backlink-pilot if	Pick LaunchRepo if
What is your budget?	Zero. You would rather spend an afternoon setting up free code than pay for a toolkit.	One payment of €49.00 is fine if it buys curated playbooks and saves you the checking.
Do you want to read and change the code?	Yes: it is MIT-licensed JavaScript you can fork, extend with adapters and run however you like.	You would rather get instructions per platform that any agent can follow, and keep the code you run small.
Volume or fit?	Volume: a list of 258 targets, 180 of them marked for automatic submission.	Fit: 299 playbooks with who each platform accepts, what gets rejected and what counts as confirmed.
How do you feel about blog comments and bot checks?	You will use the directory part only, decide for yourself on the comment module, and accept code that solves simple CAPTCHAs.	Your agent should never post comments for links or get past a CAPTCHA; it should hand those to you.
Do you launch in Chinese-language markets?	Yes: 57 of its targets are Chinese-language sites, 39 of them marked for automatic submission.	Mostly English-speaking launch platforms and directories, with playbooks written in English.

**Time-lapse · 4:21–4:52**What the run looks like: a time-lapse from the recorded demo in which Claude Code fills in listings in the founder’s signed-in Chrome.Screen recording from September 24, 2026. English audio with captions. Open the video page →

Side by side

Published pricing: free (MIT licence). Free and open source under the MIT licence: clone it, change it, run it. You run its Node.js code yourself, with bb-browser driving your own Chrome, and Claude Code can guide the setup. The price you pay is your time to run, check and maintain it.

LaunchRepo and backlink-pilot compared across twelve criteria, checked on October 1, 2026
Criterion	LaunchRepo	backlink-pilot
Pricing model	One payment. No subscription, no renewal, no fee per launch.	Free, open source (MIT licence).
What the price covers	One licence. €49.00 is today’s price as a reference incl. 19 % German VAT; checkout applies the tax for your country. The price does not rise.	Nothing to pay. You bring a machine with Node.js and Chrome, and your time to run and maintain it.
Number of products covered	Unlimited products you own and operate. The second product costs nothing extra.	One product per config file; switch the file to submit another.
Who does the work	Your own AI coding or browser agent, started by you. LaunchRepo is a toolkit, not a service.	Its CLI, run by you or by Claude Code: one site per command, or the batch module for blog comments.
Accounts used	Your own accounts, in your own logged-in browser. Nobody signs in as you.	Your real Chrome via bb-browser, with your login state; the example config keeps email and password fields for sites that need a login.
What you own afterwards	The repository with 299 playbooks, your product.json and your tracker — cloned to your machine and yours to keep.	The code, your config and a submissions.yaml history on your machine.
Directory count and source	299 directory playbooks in the repo, each with preparation, form fields, blockers and what counts as a confirmation.	targets.yaml: 258 entries (the README says 259), 180 marked for automatic submission, 45 marked dead; 4 site-specific adapters plus a generic one.
Tracking of outcomes	Your submissions.json and the local dashboard record draft, submitted, pending, scheduled, needs-you and live separately.	submissions.yaml with site, status and time, plus screenshots for manual checks.
Guarantees	None. Every directory decides for itself what it publishes, so we promise no listing, backlink or ranking.	None. MIT licence, provided “as is”.
Own AI agent required	Yes. Bring Claude Code, Codex, Cursor or a browser agent; your provider bills its usage.	Optional. The CLI runs on its own; CLAUDE.md lets Claude Code set it up and run it.
Languages	Playbooks, templates and dashboard in English; this website in 12 languages.	README in English and Chinese; 57 of the targets are Chinese-language sites.
Best for	Developers who ship more than one product and want a launch process they keep.	Developers who want free, hackable code for bulk directory submissions and will check each result themselves.

What we counted in the repository

Numbers in a README are easy to repeat. We downloaded the files at the latest commit and counted them ourselves, reading them as data without running any of the code.

The README or guide says	The file at the latest commit holds	What it means for you
“259 target sites”	258 entries in targets.yaml.	One off. The list itself is the useful part.
“226 auto-submittable with bb-browser”	180 marked auto: yes, 32 manual and 46 no; all 46 are also marked dead or paid.	Plan with 180, 137 of them English-language. The batch-comment list, a separate file, has 226 entries.
“44 confirmed dead”	45 marked dead.	About one target in six is already known to be gone, by the repository’s own labels.
“~15 paid only”	1 marked paid.	Paid status is not tracked in the file; check each site’s price before you submit.
“battle-tested on 30+ sites”	20 commits, the most recent about six months before our check; no releases.	Directory forms change often. Expect to fix adapters and targets yourself.

Counted on October 1, 2026 at commit 926d923 of the canonical repository by s87343472, the one with 366 stars and 98 forks; forks may differ. Labels such as auto or dead are the repository’s own; we did not test the sites.

Risks to check before you run it

Free code still has a price if it costs your product its reputation. These are the points we would check, with what each side does.

Risk	backlink-pilot	LaunchRepo
Blog comments for links	batch-submit.js posts 20 template comments such as “Thanks for sharing this! Really useful perspective.” under 5 rotating names, with your URL in the website field.	Not part of the toolkit. Its agent instructions say: do not manufacture comments or votes.
CAPTCHA and bot checks	Color CAPTCHAs are solved automatically, others skipped; the guide says bb-browser “bypasses all anti-bot detection, Cloudflare, and OAuth”.	Never bypassed or solved. Your agent stops, saves its work and hands the check to you, then resumes.
Passwords	The example config has email and password fields for sites that need a login, in a file excluded from Git.	No passwords in config: your agent works in the browser you are already signed into.
Double submissions	A clear rule, “Never submit the same product to the same site twice”, and a status command over the history.	Your agent searches for an existing listing first and settles any unclear attempt before trying again.
What counts as done	A status per attempt and a screenshot, which you check by hand.	Draft, submitted, pending, scheduled and live are kept apart; live needs the public page.
Code on your machine	You run its Node.js code and a global npm package that controls your Chrome with your login state.	Your agent reads Markdown playbooks; the scripts it runs, such as the tracker and the local dashboard, sit in the repository for you to read first.

From backlink-pilot’s README, guide, CLAUDE.md, example config and source at commit 926d923, read on October 1, 2026, and from LaunchRepo’s agent instructions. Google’s spam policies, linked below, list automated link creation and comment links as link spam.

Worked example: One product, a first week of submissions

Inputs — an example, swap in your own

One new product and its own logged-in Chrome, with you nearby for logins and checks.
backlink-pilot at its own pace: 5–10 sites a session, 1–3 minutes apart, from the 180 targets marked for automatic submission.
LaunchRepo: €49.00 once, your agent working through the playbooks you pick, plus your AI provider’s bill.

With backlink-pilot you pay nothing and start fast: one config file, one command per site. The work moves to checking: which of the 180 targets still accept your kind of product, which sent a confirmation, which went live, and which runs you leave out because they post comments. With LaunchRepo the checking is written down before you start: each playbook says who the platform accepts, what your agent prepares, where you step in and what proves a listing is live, and your tracker keeps those states apart. Free code versus a curated workflow: both get submissions out, but only one tells you what they are worth.

An example with stated inputs, not a measured result. backlink-pilot prices as published on October 1, 2026; your AI provider bills agent usage separately.

How many targets fit you

When backlink-pilot is the better choice

You want to pay nothing. backlink-pilot is free under the MIT licence; LaunchRepo costs €49.00.
You want code you can change. Fork it, add an adapter for a site you care about, and keep it in your own toolchain.
You want speed over curation: one command per site, with a generic adapter that tries any submission URL you give it.
You launch in Chinese-language markets: 57 of its targets are Chinese sites, an area LaunchRepo does not focus on.
You want extras in the same CLI: an issue generator for awesome lists and an IndexNow ping for new pages.

When LaunchRepo is the better choice

You want curated knowledge, not a raw list: each of the 299 playbooks says who is accepted, what gets rejected and what counts as confirmed.
You want rules your agent keeps: no comments or votes made up for links, no way around a CAPTCHA, and a clean handoff to you instead.
You want an honest record: draft, submitted, pending, scheduled, needs-you and live are separate states, and a saved draft never counts.
You want your passwords kept out of config files: your agent uses the browser you are already signed into.
You launch several products and want the process, product brief and tracker kept in one repository you own.

Verdict

backlink-pilot is a generous gift to indie developers: free code, a ready target list and a setup Claude Code can run. Use it with open eyes. Its own file marks 180 of 258 targets for automatic submission, 45 are marked dead, and its batch module posts template comments under rotating names, a pattern Google lists as link spam. LaunchRepo is the better choice when the listings carry your product’s name: €49.00 once for 299 curated playbooks, rules that keep your agent within each platform’s terms, and a tracker that tells you what really went live.

Directory submission services compared: 3 pricing models →

Decide with today’s numbers

Our pricing page loads the live price; for backlink-pilot, check their website. This page only knows what was published on October 1, 2026.

LaunchRepo pricing Visit backlink-pilot ↗

Questions about LaunchRepo vs backlink-pilot

Is backlink-pilot really free?

Yes. It is published under the MIT licence, so you may use, change and share it at no cost. You pay with your own time and machine: Node.js, a global bb-browser install, your Chrome, and checking every result.

Can backlink-pilot submit to 226 sites automatically?

Not by its own target file. The README says 226 of 259 sites are auto-submittable; in targets.yaml at the latest commit we counted 258 entries, 180 of them marked for automatic submission, 45 marked dead. Separately, its batch-comment list also holds 226 entries. Checked on October 1, 2026.

Is it safe to run?

Read it first. The code runs on your machine and controls your real Chrome with your login state, and the guide says it “bypasses all anti-bot detection, Cloudflare, and OAuth”. We read its source on GitHub and did not run it. If you try it, use a separate Chrome profile and skip the batch-comment module.

Will blog comments from backlink-pilot help my SEO?

Google’s spam policies list forum and blog comments with optimized links as link spam, and its example looks much like the module’s templates. LaunchRepo does not post comments; its instructions tell agents not to make up comments or votes.

Can I use both?

Yes, with one record. Use backlink-pilot’s generic adapter for a free site you found, and LaunchRepo’s playbooks for the platforms that matter most. Write every attempt into one tracker, so nothing is submitted twice.