Why We Built an SEO Tool for Solo Founders

ContextBolt started with a pile of saved X posts I never looked at again. Today, most of our revenue comes from an SEO tool. That second product exists because I was a solo founder doing my own SEO, and the tools I knew were built for people who do SEO all day.

On September 22, 2026, 99helpers published an interview with me in The Maker Series, their conversations with independent makers. We covered how ContextBolt is built, what worked, what didn’t, and the revenue month by month.

This post is the part of that story I care about most. Why ContextBolt SEO is built for founders, and what doing our own SEO has taught us.

It started with bookmarks I never read

I was saving useful posts on X, Reddit and LinkedIn every single day, and none of it was searchable. That was a library of knowledge I’d built up myself, and it was doing nothing.

I figured plenty of people had the same pile of saves sitting there. So we built ContextBolt Bookmarks. It’s a Chrome extension that captures what you save on X, Reddit and LinkedIn, saves any web page in one click, and tags everything with AI so you can find it again. On Pro, your AI agent can search the whole library. It went live on the Chrome Web Store in May 2026.

Building it taught me the idea behind everything we make. An AI agent is only as useful as what it can reach. It can’t see what you saved last Tuesday. It can’t see who ranks for your keyword today. Hand it that context and it gets far more useful.

The SEO tool I needed and couldn’t justify

A couple of weeks after the extension went live, I hit a different wall.

My background is growth marketing, and I’ve spent years inside keyword tools and rank trackers. Now I was doing SEO for our own product, a few times a month, squeezed between building, posting and answering support email.

The tools I knew were priced for people who live in them. Ahrefs Lite is $129 a month, and Semrush starts at $139 (list prices, September 2026). I needed a sliver of that data, now and then.

And I was already working in Claude Code all day. Every SEO question meant leaving it. Open a suite, run a report, export it, paste the numbers back into the chat. I wanted to ask “what’s this keyword’s difficulty?” and get the answer right there.

So we built that. ContextBolt SEO launched on May 20, 2026.

It’s a hosted MCP server. MCP, the Model Context Protocol, is an open standard that lets an AI agent like Claude use outside tools and data. You add one server to your agent, then ask in plain English. Keyword volume and difficulty. Who ranks for a term. What a competitor ranks for. Who links to you. Your own Search Console data. Whether ChatGPT and Google’s AI answers mention you. If you work in Claude Code too, SEO inside Claude Code walks through the setup.

The agent is where you ask, and every lookup saves automatically to your SEO Dashboard. So the research is still there next week, instead of disappearing when the chat ends.

It took me a while to see the real problem with the suites. They aren’t overpriced. They’re priced for someone else. If SEO is your job, or you bill clients for it, a suite earns its seat. If you’re a founder checking keywords twice a month, you’re paying for a cockpit to ask three questions.

What SEO looks like for a solo founder

You don’t need an SEO department. You need answers to a few questions, a few times a month.

• What am I already ranking for?
• What should I write next?
• Can a site my size actually win it?
• Did the last page work?
• Is ChatGPT mentioning us, or someone else?

Here’s the month that works for us. It takes about thirty minutes, plus the writing.

1. Read what you already rank for: Open Google Search Console and sort your queries by impressions, which is how often you showed up in search. You’re almost certainly ranking for things you never targeted. It’s free, and the numbers are Google’s own, not estimates.
2. Find your page-two keywords: Filter to positions 11 to 20. Google already thinks you’re relevant for these and doesn’t trust you enough yet. Moving one onto page one usually beats writing something new.
3. Look at who ranks before you write: Search the term and count how many of the top ten are big companies selling something. Four or more and you’re writing into a wall, whatever the difficulty score says.
4. Ship one page, then leave it alone: One good page a month beats four rushed ones. Give it six weeks before you judge it. Search moves slowly, and checking it daily only buys you anxiety.

With ContextBolt SEO and your Google Search Console connected, steps two and three become one message. Our guide to striking distance keywords covers step two in depth.

Copy this prompt

Pull my Search Console queries from the last 90 days that rank
between positions 11 and 20. For the five with the most
impressions, show who holds the top three spots and the keyword
difficulty. Tell me which one I can realistically move first.

The Search Console part is free. The results-page and difficulty checks cost one credit each. We wrote the whole routine up on our page about SEO for solo founders, including what’s free, what’s worth paying for, and when a suite really is the right call.

What building ContextBolt taught us

Some of this I’d tell any founder. Some of it I’m still learning.

Social got us started: A few posts went semi-viral early on, and that’s where our first paying users came from. Social has been quieter since, and search has picked up the slack.

The free trial changed the most: Until July 22, 2026, you had to pay to try ContextBolt SEO. Then we added a 7-day free trial with 100 credits. People could see the value before paying, instead of taking my word for it. It’s helped conversions more than anything else we’ve changed.

Focus beats range: Since launching, I’ve started a few other things, including a meeting coaching tool that got benched early. The lesson stuck. Pick one thing and stick with it.

Distribution is the real job: I said this to 99helpers, and I’d say it again.

Building the product is the fun part, and honestly the easy part. Sitting down and doing the distribution work every day is much harder, and it’s the thing that actually grows the business.

Celebrate the small wins: A new subscriber or a kind message from a customer is what keeps you going early on. The comment I hear most is that ContextBolt SEO doesn’t feel like a whole new system to learn. People add it to their agent and start asking. It’s easy to skip past moments like that and only stare at what isn’t working yet.

We grow ContextBolt with ContextBolt SEO

We use our own tool for our own SEO. It’s a nice loop to be in.

Our blog posts usually start as a question to ContextBolt SEO. Is anyone searching for this? Who ranks now? Could a site our age win it? If the answer is no, the post doesn’t get written. Our guide to keyword research with Claude walks through the same checks.

It’s working. From June 11 to September 10, 2026, Google Search Console counted 5,483 clicks to contextbolt.com, and we showed up in search 886,308 times. That’s about 60 clicks a day, and we didn’t pay for any of them.

Two things we learned along the way.

• Write for the problem, not the tool: From June 4 to August 31, 2026, our posts that solve a real job, like exporting Reddit saves, got a click through to the product on about 4% of views. Setup guides for other companies’ tools managed under half a percent.
• Start in month one: SEO takes months to pay back. That’s the reason to start early, not the reason to wait.

Advice for founders starting now

This is what I told 99helpers, and it’s what I’d tell a friend starting today.

1. Start small: Pick a problem you have yourself, or one people keep complaining about on Reddit, and solve that first.
2. Get people using it early: Ask for feedback while it’s rough. Early on, it’s worth more than your roadmap.
3. Keep showing up: Post every day. Start your SEO early. Both compound.
4. Double down on what people pay for: Once something earns money, push on that one thing. Try not to spend time on anything that doesn’t bring money in.

Right now you can build almost anything. Most of ContextBolt is built with Claude Code and Codex. The hard part left is getting it in front of people.

Where ContextBolt goes next

The goal hasn’t changed. Build the best SEO tool we can for founders who won’t pay for, or learn, the big enterprise suites. Make SEO something a one-person company can actually do.

The number we’re chasing is $10,000 a month, bootstrapped, with no funding. When 99helpers asked in September, we were at $653 a month, up from $47 at launch in May. Small, but it’s moved the right way every month.

If you’re a founder doing your own SEO, try ContextBolt SEO free for 7 days. If you want the whole story, with the stack and the numbers month by month, the 99helpers interview has it.

There’s a lot of distribution between $653 and $10,000. I’d better get back to it.