Voice AI for Inbound Calls
Answers the phone during the dinner rush, takes reservations, handles questions. Runs on your Twilio number and Vapi account. You own the code.
$2,500 setup + $400/mo managed
Restaurant page
Why no client list — and why that's the honest answer
Baxter Solutions is new. This is month one. There is no client list, no testimonials, and no logos on a "trusted by" wall — and I'm not going to manufacture those.
What I can offer instead: published prices you can evaluate before we ever speak, a methodology you can read on this site, and the fact that I personally build everything I quote. There's no team delivering work I've promised. It's me, which means you get the person you talked to — not a project manager and a junior developer you've never met.
The market I'm entering is real and validated. Hostie raised $4M from Gradient Ventures (Google's AI fund) to build restaurant voice AI. Loman raised $3.5M and published a 22% revenue lift study for pizza operators. AutomateNexus has published a $7,500 BYOK workflow build price in San Diego. These are verifiable companies solving verified problems — I'm not inventing the demand.
What I'm offering is the same delivery model — client-owned, client-hosted, bring-your-own-keys — with published prices, a specific audit that protects you from committing to a build that won't pencil out, and someone who will show up in person in San Diego if you want to see it running before you sign anything.
If you want the large firm with a portfolio, I can point you toward one. If you want fixed prices, a written ROI analysis, and code that keeps running after I'm done — that's what this is.
Why I build systems you own instead of systems I rent you
Most automation agencies build on their infrastructure. When you cancel, the system goes with them. This is the standard model — and it's fine if you understand the tradeoff.
I don't do that for a few reasons that are worth being direct about.
First, it's better for clients. A system that runs on your accounts and your servers is a system you control. When I raise prices, you're not stuck. When I'm busy, your phone still gets answered. When you want to move on, you can take the code to any developer. The system is an asset you own, not a subscription you rent.
Second, it's the right product for independent businesses. Restaurants, small practices, and lean SMBs have already been burned by vendor lock-in. A Clover POS that doesn't export your data. A CRM you can't get out of. A marketing platform that owns your contact list. The "you own it" pitch is not a feature — it's a relief. Every client I've talked to responds to it the same way.
Third, the open-source stack I build on — n8n, primarily — is explicitly designed for this model. The fair-code license requires client-hosted delivery for any commercial multi-tenant deployment. This isn't a workaround; it's what the stack was built for.
The tradeoff: you need a hosting account (Hetzner, Railway, or Render — typically $10–$50/month) and your own API keys with the relevant vendors. It's one setup step. After that, you own a working system and nobody can take it away by changing a pricing page.
Why every price is on the site before you ask
Every automation agency in San Diego I could find either doesn't publish prices or has a "contact us for a quote" page where you trade your email for a sales call. I don't do that.
Published prices do two things. They protect you — you can evaluate the math before we ever speak. And they filter the work — if a prospect can't see the value in a $1,500 audit, the build is going to be painful for both of us. The price on the page is the qualifying mechanism, not the starting point for a negotiation.
The audit ($1,500 standard, $2,500 for restaurant or VA-replacement work) produces a written report with your actual numbers — your call volume, your VA spend, your time investment — before you commit to a build. The build is quoted inside that report. Nothing gets re-priced after you've already agreed to proceed.
If your situation genuinely requires a custom quote — multi-location, three integrated systems, real compliance constraints — the audit surfaces that and we quote accordingly. But the published ladder covers the realistic range for what small businesses actually need.
See the full price ladder on the homepage pricing section. Every offer, every tier, every number — on one page, before you ask.
What I bring to a build
UCSD data science. Built working systems before this agency existed. This isn't a pitch — it's what I can actually do.
Data Science — UCSD
Systems thinking, not just tools
Data science trains you to ask what the data actually shows before you build the model. That same instinct applies to automation: we figure out what the actual problem costs before we quote a fix. The Operations Audit is that step formalized.
Built things
Systems that ran before I started selling them
Before opening Baxter Solutions, I built multi-agent automation pipelines, business data processing tools, and workflow systems as personal projects. The methodology on this site came from building things, not from a course.
SD-local
Can show up in person
North Park, Little Italy, Mission Valley, La Jolla, Pacific Beach — that's a drive, not a flight. Walk-in demos are a real option for any San Diego restaurant or business that wants to see the system running before committing to anything.
What I don't have (and what that means for you)
The honest version of my credentials
No client case studies yet. This is month one. I don't have before-and-after numbers from a restaurant client, because there are no restaurant clients. The proof points on this site — Hostie's funding, Loman's 22% lift, AutomateNexus's $7,500 price — are market facts I can back with sources, not client outcomes.
No large team. It's me. That means no PM layer, no outsourced delivery, and no gap between what I quoted and what gets built. It also means my capacity is limited — I won't take on more work than I can actually deliver.
No enterprise security infrastructure. SOC 2, HIPAA BAA, FedRAMP — not yet. If you're in a compliance-gated context, I'll tell you that clearly before we spend time on discovery. This is a constraint, not something I'll work around quietly.
The tradeoff: you get the person who built the thing, published the prices, and can walk through your restaurant on a Tuesday. The large firm with the client list charges more and delivers through someone you haven't met. Both are valid. I'm telling you the actual difference.
Three problems, three pages
We focus on what we build well and say no to the rest. Outside these three, we'll tell you honestly if we're the right fit.
Replace Repetitive Offshore Work
Map your VA task list to an n8n workflow. Audit produces a 12-month TCO comparison. Build from $7,500. You own the code.
$2,500 audit · build from $7,500
VA replacement page
Connect the Tools You Already Pay For
Integration between Weave, NexHealth, and OpenDental at the scheduling layer. We do not touch EHR clinical data — scheduling, recall, and intake only.
$1,500 audit · build from $7,500
See on homepage
Reach out directly.
No form required.
Email is the fastest path. One sentence about what you're trying to fix — that's enough to start a conversation.
- ethan@baxter.solutions
- San Diego, CA — on-site demos available
Response time: within 1 business day, usually faster. If you're a San Diego restaurant or business with a specific operational problem, describe it in two sentences — that's enough.
Send a Message
Or email directly — either works