Howitzer Reviewthe first Reddit DM tool, and how it disappeared

Howitzer was a Reddit outreach tool that described itself, in its own meta description, as “the first Reddit marketing automation tool”. It was aimed at founders, agencies and growth marketers, and its archived homepage summarised the pitch in a single line: “Find new customers on Reddit, and send them personalized messages.”

Mechanically it had two halves. The first built a list. You wrote what its documentation called a query — get all users who posted or commented using this keyword, in these subreddits, above this upvote score, in this time window, with positive or negative sentiment — and Howitzer returned a list of usernames with the matching post attached. Lists could be combined, intersected, imported and exported as CSV, and you could pull every active member of a subreddit.

The second half messaged that list. Howitzer sent Reddit private messages with personalisation variables for username, matched keyword and subreddit, ran a built-in spam checker over your draft, handled follow-up sequences, and collected everything in a Facebook-style inbox so you never left its dashboard. Its plans included sending accounts: “We send the messages for you, through our high-karma Reddit accounts”, as the archived homepage put it, though you could connect your own instead. Two features were listed as coming soon in 2022 rather than already shipped: “campaigns on autopilot”, which would watch subreddits in real time and message people seconds after they posted something relevant, and “Reddit Chat Automation”.

The company behind it was not a weekend project. Its own FAQ names three founders — Nikola Velkovski, Stefan Kotevski and Tomislav Jurukovski — a team of eleven, and a $500,000 pre-seed round led by Day One Capital of Budapest. It launched to #1 Product of the Day on Product Hunt, a result its CEO wrote up on Indie Hackers in April 2022.

The testimonials Howitzer chose to publish are a decent guide to what the product felt like to use, even allowing for the obvious selection bias in a vendor's own wall of quotes.

The recurring theme is reach that felt disproportionate to effort. A user quoted on the archived homepage reports “38 responses from 200 messages”; another describes it as “getting 3x opt-in rate compared to email marketing”. Howitzer's own headline statistics — self-reported, and not independently verified anywhere — claimed a 41% average response rate and a 5.5% conversion rate across its campaigns, alongside 870,000+ messages sent and, depending on which capture you read, between 950 and 1,450 customers.

The second theme is that people used it for things other than selling. One quoted user says they used it “for growing my subreddits and conducting user research”; another used it to find people who had mentioned Product Hunt and invite them to a launch. Whatever you think of the ethics of the mechanism, the targeting engine underneath was good enough that people repurposed it.

The third theme, visible in the same testimonials, is that experienced users treated it as something to handle carefully. Buyers on lifetime-deal marketplaces recommended going through the training material “so you don't risk your reddit account getting banned”, and Howitzer's own FAQ advised “sending no more than 50 messages per day per account”. The people who liked it most were the ones who already understood they were operating close to a line.

Sentiment as a targeting axis was the cleverest thing in the product, and it has not really been improved on since. Being able to say “find me people who posted negatively about this competitor in these five subreddits last month” is a materially better filter than a keyword alone, and Howitzer had it as a plain checkbox in the plans it published in 2022.

The list algebra was the second thing it got right. Lists that could be combined and intersected turned an ad-hoc search tool into something you could build a repeatable process on, and CSV import and export meant your work was portable — a quality worth more than it looks, given how the story ended.

It was also unusually honest about the risk it was asking you to take. A built-in spam checker, an explicit daily send recommendation, and an FAQ that told you plainly “you cannot expect to sell through a single cold message sent to someone you don't know” are not the marks of a tool trying to hide what it does. It also stated, in the same FAQ, that it was “not an official feature of Reddit Inc.” and not connected to Reddit in any way.

And the launch execution was excellent. #1 Product of the Day, a founder who wrote up the playbook publicly afterwards, a Discord community, a visible presence on Indie Hackers — the go-to-market was as considered as the engineering.

The strategic flaw is the one everything else follows from: the product's core action was the specific behaviour its host platform prohibits. Reddit's spam policy names “sending large amounts of unsolicited chat or private messages” as an example of what may violate it, and defines spam to include actions that are “repeated or unsolicited” whether automated or manual. There was no version of Howitzer's roadmap that made that tension smaller.

Supplying sending accounts made it sharper still. A tool that sends messages “through our high-karma Reddit accounts” is running the risk centrally, which is efficient right up until the moment it isn't. It also meant customers' outreach was partly identified with infrastructure they did not control.

There was a narrower product gap too. Howitzer was outbound-only: it never did anything with the inbound half of Reddit, where people publicly ask which tool to use and would welcome an answer. The engine that found “users who mentioned X” was one small step from “threads worth replying to”, and it never took it. In our opinion that is the missed pivot, and it is a fair part of why products like ours exist.

Finally, the wind-down was handled poorly for customers, which is the thing former users mention most. There was no closure announcement addressed to Reddit customers, no export deadline, no migration guidance — the messaging feature simply “moved to private” in an FAQ answer, the product site eventually stopped resolving, and the domain was pointed at a LinkedIn tool. Compare that with GummySearch, which gave three weeks' notice and a long data-retention window when it closed in 2025.

Because there was no announcement, the timeline has to be assembled from the Internet Archive. It is worth laying out precisely, because most write-ups on the web give a single wrong date.

March and April 2023: the Reddit product site carries a banner announcing HeyReach, the team's new LinkedIn tool, and its FAQ answers “What happened to the messaging part of Howitzer?” with “Howitzer's messaging feature has moved to private”, explaining that the team wanted to focus on its “bigger vision”. The same answer describes two products in development, a LinkedIn automation tool and a multi-platform prospecting tool. Lead extraction survives; the messaging does not.

Mid-2023: co-founder Nikola Velkovski later told Swiss EP that “by mid-2023, Reddit's API crackdown, user backlash, and monetization headwinds left the team on the verge of collapse”, and that “in July 2023, we were literally two months away from bankruptcy”. This is the same period in which developers of major third-party Reddit apps publicly said Reddit's new API pricing made their products unviable.

December 2023 onward: the Reddit product site returns a server error in archive captures, and still did when we checked it in July 2026.

September 2024: howitzer.co is a company page with two products — HeyReach for LinkedIn, and “Howitzer for Reddit”, now described as a way to find leads and “manually reach them through chat”. January and February 2025: the domain is captured first as a 404 and then as a redirect to heyreach.io. That redirect is still live in July 2026. The Reddit product's public life ran roughly from 2021 to early 2023; the brand's took two more years to finish unwinding.

First, ask what a tool's core action looks like against the host platform's own rules — not its terms of service in the abstract, but the specific behaviour you will be doing hundreds of times a day. If the answer is a rule the platform has written down and named, price that in. It is not a reason never to buy, but it is a reason not to build a pipeline that has no second leg.

Second, treat a vendor's data supply as part of the product. Both Reddit tool shutdowns of the last three years — Howitzer's and GummySearch's — trace back to Reddit's data terms rather than to a lack of customers. Multi-source tools degrade when one platform changes; single-source tools end.

Third, judge vendors on how they behave when they leave. Howitzer's Reddit customers got no notice, no export window and no forwarding address, and that is a more useful signal about a company than any feature comparison. Ask, before you commit to an annual plan, what happens to your data if the product is discontinued, and whether you can export it yourself today without asking support.

For what it is worth, the parts of Howitzer worth missing are the targeting parts: sentiment as a filter, list algebra, a real query language over Reddit. Those ideas deserve to outlive the delivery mechanism they were attached to.