I Rejected a $15K Acquisition Offer for My Multi-Agent IDE — Here's the Full Breakdown

Last month, a CTO offered me $15,000 to acquire 1DevTool, my solo-built desktop app for AI coding agents.

I said no.

This isn't a "follow your passion" post. It's a detailed breakdown of the exact valuation math, the decision framework I used, and the concrete moves I'm making instead — so if you face a similar moment, you have something more useful than vibes.

---

What is 1DevTool, and What Were the Numbers?

1DevTool is a local-first desktop app that lets developers run multiple AI coding agents — Claude Code, Codex CLI, Gemini CLI, Aider, Cline, local models — from one unified GUI instead of juggling 6 terminal windows across 4 monitors.

Core positioning: neutral aggregator, one-time purchase, BYOK, data stays on your machine.

Pricing tiers at time of offer: $29 / $69 / $89 (lifetime).

Metrics at the time of the offer:

• MRR: ~$300/mo

• Annualized revenue: ~$3,600/yr

• Active users: ~20

• Users in close feedback loop: ~10

• International sales (zero intl. marketing): ~3/month

• Primary channel: Vietnamese dev communities

• Age of product: < 12 months

---

The Offer: Was It Fair?

The buyer — I'll call him Marcus — was a CTO who had actually used the product. He filed real bugs, gave real feedback, and made a clean cash offer: $15,000 at close, full acquisition of code and customer list.

Valuation math:

• $15K ÷ $3,600/yr ARR = 4.17x ARR multiple

• Typical range for one-time-purchase desktop software on Acquire.com or MicroAcquire: 1.5x–3x ARR

• So on a pure multiple basis, this was a fair-to-good offer

But ARR multiples are the wrong lens for an early-stage product with a non-linear growth story. The right question isn't "is this a fair price for today?" It's "what does today's price imply about the ceiling?"

At $15K, Marcus was implying a terminal value roughly equivalent to 4 years of flat $3.6K/yr revenue — with no growth, no upside, and a modest risk discount. That's the bear case priced as the base case.

---

Why I Almost Said Yes: The Fear Inventory

I want to be honest about what was actually scaring me, because I think most "sell vs. keep" posts skip this.

Fear 1: The big labs are shipping their own IDEs — and they're good

OpenAI launched Codex with native computer use built in. That's not a plugin or a third-party integration — it's first-party, deeply integrated, and ships with the full weight of OpenAI's distribution behind it. Anthropic, Google, and others are moving in the same direction. When a lab ships their own IDE, they're not just a model provider anymore — they're a direct competitor with zero marginal cost to bundle features I charge for.

Fear 2: Competitive intensity is accelerating

Cursor, Windsurf, Claude Squad, and others are all well-funded and shipping fast. I'm one founder in Vietnam.

Fear 3: I was spread thin

At the time I had multiple products in flight — 1DevTool, Server Compass, plus content properties. Selling one felt like buying focus.

These were my reasons to sell. They were real. I wrote them down. Then I wrote the other list.

---

Why I Said No: The Conviction Inventory (With the Actual Reasoning)

1. The neutral aggregator moat is structural, not temporary

This is the insight that mattered most.

Anthropic optimizes for Claude adoption. OpenAI optimizes for Codex stickiness. Cursor has a financial relationship with Anthropic. None of them have incentive to build a layer that works equally well across competing models.

But power developers don't monoculture their stack. They use Claude for long-context reasoning, Codex for fast completions, Gemini for multimodal, local models for privacy-sensitive work. The AI coding stack is fragmenting, not consolidating, at the model layer — even if individual vendors want it to consolidate at the client layer.

The insight: every new AI CLI that ships is a growth event for 1DevTool, not a threat. Each new tool means one more terminal window that needs to live somewhere.

This structural advantage doesn't depend on beating Anthropic at features. It depends on staying neutral when everyone else is choosing sides.

2. The Vietnam/SEA signal was being misread as a weakness

I had ~3 international sales/month with zero international marketing effort. I was reading this as a low ceiling.

It's the opposite signal.

Anthropic won't write Vietnamese content. They won't engage Vietnamese dev Discord servers. They won't show up in SEA-specific dev communities. The 3 organic foreign sales were happening despite zero supply-side effort — meaning there's demand with no competition for it.

The correct read: There's a market segment that first-party tools structurally cannot serve. I can. That's a durable niche, not a limiting one.

To validate this, I mapped where my organic traffic was actually coming from. The result: developers in Vietnam, Indonesia, and the Philippines — all markets where AI coding tools are popular but localized content is nearly nonexistent.

3. SEO compounds; selling means giving away compounding

I've been building content across three properties with consistent cross-linking. None of those properties were earning meaningfully at the time of the offer. But SEO doesn't pay linearly — it pays on a delay curve with a hockey stick tail.

The keywords I'm targeting aren't high-volume generic terms. They're intent-specific, low-competition terms that larger players ignore:

• claude code GUI — developers who know the tool but hate the terminal experience

• multi agent IDE desktop — the workflow, not the brand

• BYOK AI workspace — the pricing model as the differentiator

• run Claude Code and Codex together — the exact problem 1DevTool solves

Selling at month 8 would mean handing a compounding SEO asset to a buyer who'd extract the value I planted. That's the equivalent of selling a stock the day before it vests.

4. One-time pricing is a genuine differentiator at this specific moment

Subscription fatigue is real and growing. A developer who runs Claude Max ($100/mo) + GitHub Copilot ($19/mo) + Cursor Pro ($20/mo) is already paying $1,600+/year in AI tooling subscriptions.

1DevTool at $29 one-time

I also believe I'm underpriced. The fact that Marcus — a CTO — offered $15K for the whole company suggests the per-seat value is significantly higher than my pricing.

5. The downside of keeping is bounded

This is the most underrated reason.

Keeping: I use 1DevTool every day. It is my daily driver. The marginal maintenance cost is near-zero because I'm using it regardless. Worst case is I ship no new features and it generates $0 — no capital lost, no time lost beyond what I'd spend on my own workflow anyway.

Selling: I give away all upside. I lose my daily workflow tool. I accept a price that assumes the bear case. And it doesn't even solve the competitive fear — that risk just transfers to the buyer. I'm paying someone else to hold my upside.

The asymmetry is obvious when stated this way. Keeping has a bounded downside. Selling has an unbounded downside (what if it was worth $200K in 18 months?).

---

The Reframe: This Was a Focus Decision, Not a Sell Decision

When I put the two lists side by side:

• 3 fear-based reasons to sell (features copyable, cost headwinds, competitive pressure)

• 5 conviction-based reasons to keep (structural moat, untapped SEA market, SEO compounding, pricing position, bounded downside)

The honest conclusion was: I was considering selling a growing asset to solve an attention problem. Those are different problems with different solutions.

The right fix for "I'm spread too thin" is to cut ambition from other products, not cut the growing one.

---

What I Did Instead: The Actual Decisions

The 5 Rules I Locked In

Rule 1: Floor is $15K — but only for a focus decision, never a fear reflex.

I won't sell below $15K. And even above $15K, I only sell if it's a deliberate strategic reallocation — not because I'm scared of Anthropic's roadmap.

Rule 2: Convert the buyer into a power user.

I sent Marcus a free lifetime license and told him I'm holding and growing, revisit in 6 months. This does three things: (a) keeps the optionality alive, (b) turns a potential future buyer into an active feedback source, (c) costs me nothing. Free option on a future $100K exit.

Rule 3: Let 1DevTool ride on daily-use improvements. Reallocate marketing to Server Compass.

Server Compass has less competition and a cleaner niche. I'm shifting active marketing effort there while 1DevTool grows organically through SEO and product quality.

Rule 4: Freeze scope.

I removed the database client from the roadmap. Fewer features = less maintenance surface = lower solo-founder risk. Depth in the multi-agent niche beats breadth into territory that Cursor and Windsurf own.

Rule 5: Pre-define the shelf trigger.

If organic revenue decays 3+ consecutive months AND I stop using it daily, I'll sell or shelve — on my terms, with a real data room, not from anxiety. The trigger has to be pre-defined, or anxiety makes it for you.

---

The Honest Takeaway

$15K is 4 months of a decent dev salary in Vietnam. It's not life-changing. The regret of handing away a daily-driver project that anchors your professional credibility — and watching someone else extract the value you built — that compounds in the wrong direction.

My goal is a $10K/mo product portfolio. 1DevTool is one tile in that mosaic. You don't sell a growing tile to buy focus. You cut the tiles that aren't growing and let the good ones ride.

The hardest part of this decision wasn't the analysis. It was being honest about whether I was reasoning from conviction or from fear. Once I separated those two lists, the answer was obvious.

One ask: If you're a solo founder who has faced (or is facing) a sell/keep decision, I'd genuinely like to hear how you ran the analysis. The more data points on this, the better the framework gets.

---

Building 1DevTool — a neutral desktop hub for multi-model AI development. One-time purchase, local-first 1devtool.com

Follow the build-in-public journey on Indie Hackers or X https://x.com/khoa_solo.