Why we built FishOut
There are already a lot of social listening tools. We counted ten serious ones before we wrote a line of code, and we still track them — their pricing lives in a dated dataset we re-check rather than in a slide someone made once.
So the reasonable question is why anyone should build another. The answer is not that the existing ones are bad. Several are very good, and we say so on their comparison pages. It is that the category was designed for a job that has quietly changed.
The job used to be a report
Social listening grew up as a marketing function. The output was a report: share of voice, sentiment over time, a spike you could point at in a meeting. The buyer was a comms team, the budget was annual, and the interface was a dashboard somebody opened on a Tuesday.
You can still see that shape in the prices. Of the ten tools we track, the entry point for watching a serious number of platforms is $199/mo at Octolens, $249/mo at Brand24, and $249 for listening at Sprout Social on top of $99 per seat. One does not publish a price at all — Mention's entry is "contact sales."
Those are not unreasonable prices for what they are. A team of five people producing monthly reports for executives gets its money back. But it tells you who the product was drawn for.
The job most people actually have now is smaller and more urgent than a report. Someone is asking which tool to buy, in a thread, right now, and either you are in that conversation within a few hours or you are not in it at all. That is not a reporting problem. It is closer to an inbox.
Three free tiers out of ten
Here is the number that decided it for us. Of the ten tools we track, three have a permanent free tier: F5Bot, KWatch.io, and Google Alerts. Of those three, one watches Reddit, Hacker News and Lobsters; one watches six platforms; and one is Google's own index rather than social at all.
Everyone else starts at $29.95 and goes up steeply.
That means the person who most needs this — someone with a product, no budget, and a real need to know when their name comes up — has essentially two options, both of which stop at a handful of sources. Everyone else is asking for a card before you have seen a single mention.
We think that is the wrong shape, and not for charitable reasons. This is a product whose value is obvious in a week and impossible to explain in a paragraph. Charging before someone has seen their own mentions is a worse business, not a more generous one.
The part that actually made us start
The specific thing we could not get past was where the API sits.
Look at what it costs to get your own mentions out of these products programmatically. F5Bot's API is on Platinum, at $214.99/mo — the tier above the one most people would ever reach. Brand24 charges an additional $149 fee for REST access on its Pro and Business plans, and includes it only on Enterprise, which starts from $1,499. Awario puts its API on Enterprise at $399/mo, and Sprout Social gates its own to Advanced at $499 per seat. And when we surveyed twelve tools for an earlier post, nine of them shipped an MCP server — but the cheapest plan carrying one ran from $49 to several hundred dollars a seat, and one of the nine would only quote on request.
Your mentions are your data. Somebody wrote about your company in public. The idea that reading it back through your own code is a premium feature is a pricing decision dressed as an architecture.
So on FishOut the REST API, the MCP server and webhooks are on the free plan. Not a trial of them. The actual thing.
"AI-native" is a claim about design order, not a feature list
We want to be careful here, because this phrase has been worn out by people who added a chat box.
We are not the only company thinking about agents. Octolens publishes a 23-tool MCP server on every plan and documents it better than most. Brand24 includes MCP on every tier. Anyone claiming to be alone in this category is not counting.
The difference is order of operations, and it shows up in the shape of the thing rather than in a feature checklist.
If you have a five-year-old REST API and an agent surface arrives late, the cheapest way to ship one is a wrapper: one tool per endpoint, generated. It works, it demos, and it inherits a problem — every tool definition is injected into the model's context on every request, whether it is used or not. Anthropic's own guidance names this exact failure: "a common error we've observed is tools that merely wrap existing software functionality or API endpoints."
We did not have an existing API to wrap. The agent was the design brief, so the tool surface is small on purpose — search mentions, read one, manage keywords, act on one — and the scopes are the same objects the REST API uses, because they were designed once rather than twice.
That is a real difference and it is also a modest one. It does not make our triage better. It means that when you point an assistant at your mentions, you are not paying context rent on forty generated tools to ask one question.
What we will not do
A manifesto that only lists advantages is an advert. Three things we have decided against, which constrain the product:
We draft replies. We never post them. Automated posting and vote manipulation break the rules on most of these platforms, they are bad for the communities the product depends on, and they are a fast route to a ban that takes your brand with it. The draft opens with the thread loaded and your voice selected; you send it, from your account, as yourself.
We will not sell you an AI visibility score. There is a whole category doing this now. When we went through the research on AI citations, the reproducibility numbers were bad enough — under a 1 in 100 chance of getting the same list of brands twice — that we do not believe most of those numbers mean anything. We would rather watch the conversations those systems read than sell you a metric we would not trust ourselves.
We will not put the exit behind a sales call. Everything is self-serve, the pricing is on the page, and your data comes out through the same API you put nothing into.
Where we are honestly behind
Since we are asking you to believe the rest of this:
Our free plan covers the open sources — Reddit, Hacker News, Bluesky, GitHub and Google News — with 10 keywords and 500 relevant mentions a month. YouTube, Medium, LinkedIn, TikTok and X start at $19/mo, and that is not free.
AI triage runs on every plan, Free included, but reply drafts start at $19/mo. And the free allowance is small on purpose: it is enough to see whether the feed earns its place, not to run a category on.
And F5Bot at $9.99 for twenty keywords is below anything we sell. If cost is the deciding factor and Reddit is where your buyers are, that is the honest answer and we say so on its comparison page too.
We are also web-only. There is no desktop app, no mobile app, and nothing runs on your machine.
What we are actually betting on
That the interesting unit of work is one conversation, not one month.
Everything follows from that. Mentions arrive in an inbox rather than a dashboard, because an inbox implies somebody is meant to do something. Every mention carries a relevance with the reasons attached, because a ranked list is only useful if you can tell when it is wrong. Assignment and status exist because the second person on your team is the point at which email stops working. And the whole thing is readable by your own code on the free plan, because the mention is the start of a workflow rather than the end of one.
If that is the shape of your problem, the free plan is 10 keywords and takes about two minutes, and there is no card. If it isn't, several of the tools on our comparison pages are genuinely good and we will tell you which one.
Competitor pricing here comes from our own tracked dataset, read from each vendor's pricing page on 5 September 2026 and re-checked quarterly. If we have something wrong about your product, tell us and we will correct it — we have already corrected four prices this year.