Comparison
Tunovix vs. Xerx
Both are EU-based, fixed-price AI studios (Xerx: Innsbruck, Austria; Tunovix: Latvia) — a genuinely comparable category, unlike the DIY-platform comparisons on this site. Xerx explicitly advertises RAG and evals; Tunovix publishes a live, checkable MCP server and RAG search instead of describing them.
| Dimension | Tunovix | Xerx |
|---|---|---|
| Price | $1,150–$12,050 per build, flat $50/hr | From €4,900 fixed; €1,490 fixed for an "AI potential analysis" first |
| Source-code / IP ownership | Full ownership, delivered to your repo | Not publicly stated on their site |
| MCP support | Yes — live public MCP server at tunovix.com/mcp | Not publicly stated on their site |
| RAG support | Yes — live at tunovix.com/api/ai-search | Yes — explicitly advertised ("wired in with RAG, evals, and cost controls") |
| Integrations | CRM, helpdesk, ERP, Slack, Gmail, PostgreSQL, custom APIs | Claude, GPT, Gemini, Vercel AI SDK, n8n, Firebase |
| Deployment | Your infrastructure/repo, no vendor lock-in | Not specified on their site |
| Support model | Founder-led, one point of contact, no account-manager layer | Ongoing tuning mentioned ("we stay close to keep the system tuned"); no formal SLA stated |
| Typical timeline | 2–12 weeks depending on scope | 2–4 weeks, scope capped |
Xerx figures sourced from their own public services page as of August 2026 — check their site directly for current numbers. "Not publicly stated" means absent from their page, not confirmed absent from their actual service.
Where Xerx is ahead on paper
Xerx names RAG and evals directly in their own marketing copy — a more explicit technical pitch than Horizon's, and worth taking at face value even though the specifics (which eval framework, how RAG is implemented) aren't public. A fixed-price "AI potential analysis" at €1,490 is also a genuinely useful lower-commitment first step if the actual question is still "should we build this at all."
Where the two studios differ in practice, not just claims
Tunovix's MCP server and RAG search are both things you can query right now — tunovix.com/mcp and tunovix.com/api/ai-search — rather than a description to take on trust. See MCP development and RAG development for what's actually behind those. Neither studio publishes source-code ownership terms, deployment specifics, or a formal SLA — worth asking directly, on either side, before committing.
What actually matters for your decision
Same advice as every comparison on this site: ask for real, contactable references, not testimonials. See Tunovix's own named case studies, or the fuller EU buyer's guide for what else to check before choosing between fixed-price studios.
Have a product in mind? Let's scope it this week.
A 30-minute call gets you a rough timeline and cost — no obligation, no sales deck.