Accounting01
Maria
15 hours a month manually logging supplier invoices, categorizing expenses, preparing the monthly report. Same operations, same accounting software, every time.
Flagship product · Studio Futuro
Spaceship, software that evolves with your company.
Every company needs a different spaceship. That's why we start from a deep analysis of your processes — we figure out where AI actually makes sense and where it doesn't — then we build the modules around you and crew them with AI astronauts pre-trained on your business.
· Desktop · macOS + Windows· Analysis + custom configuration · on-premise or cloud
The problem · Repetitive work
01 / 08Operators lose hours on the same tasks, every week.
It's not a personal productivity problem. It's that generic tools — ChatGPT, Copilot, Zapier — don't know the craft. They ask people to learn prompt engineering when those people don't even have time to empty their inbox.
Sales02
Luca
Answers the same twenty emails a day: quote requests, deal follow-ups, order status updates. The CRM gets updated when he remembers.
Marketing03
Sara
Plans content, analyses campaigns, digs for performance data across analytics tools. Spends more time collecting data than deciding what to do with it.
Admin04
Giovanni
Organises documents, schedules meetings, tracks contract deadlines. Compliance and order done by hand, in a spreadsheet only he understands.
How it works · For non-technical users
02 / 08Like hiring a colleague. Just digital.
Six steps, zero prompt engineering. Below is the actual flow Maria lives every day since Spaceship went live in her company.
- 01Step / 01
Maria's problem
Maria works in accounting. Every month she spends hours manually logging invoices, answering the same emails, hunting for documents. Repetitive work, no time for anything else.
- 02Step / 02
She opens the app
She installs Spaceship on her computer. It's a normal app, like Slack or Outlook. No technical setup, no code, no 'prompts'.
- 03Step / 03
She logs in
She signs in with her corporate email (or Google). She's in within 5 seconds. Same security as the company website.
- 04Step / 04
She asks in plain language
She writes to the assistant like to a colleague: 'Log the invoice from supplier X and send me a copy by email.' Nothing technical.
- 05Step / 05
The assistant works
Spaceship understands and connects to her tools: accounting software, email, bank, drive. The answer appears word by word, live.
- 06Step / 06
Maria approves
Before touching anything important, the assistant asks for confirmation. Maria clicks 'OK'. Job done. Hours saved, zero surprises.
Modules · Four crafts, one workspace
03 / 08Studio Futuro activates the modules your company actually needs.
Each module is a pre-trained agent connected to the tools of that craft. You only pay for active modules, and add more when they're useful.
Routing per processo
Available
Accounting
Invoice listing and search, expense categorization with human approval, aggregated monthly reports. Connects to the accounting software, the bank, spreadsheets.
Coming soon
Sales
Pipeline tracking, quote generation, automated follow-ups on open deals. Connects to the CRM, email, calendar.
Coming soon
Marketing
Content planning, campaign analysis, social scheduling. Connects to analytics, social APIs and the company CMS.
Coming soon
Administration
Document organization, meeting scheduling, deadline and compliance tracking. Connects to file storage, calendar, email.
Why Spaceship · Three real differences
04 / 08Three things generic tools don't do.
This isn't a ChatGPT wrapper. Spaceship is built around three technical decisions that change how AI enters a company.
01Pillar
Pre-trained on the craft
Modules arrive pre-configured on the department's flows. The accounting agent knows what 'categorize an invoice' means, it knows the typical chart of accounts, it speaks technical Italian. The operator doesn't write prompts: they write the way they'd talk to a colleague.
Tech · Mastra persona + dedicated MCP tools · Claude 3.5 Haiku · typed Zod schemas
02Pillar
Human-in-the-loop by default
Every write operation — categorization, email send, CRM update — pauses and asks for explicit approval. The operator sees a card with what's about to happen, clicks Accept or Reject, the work continues. Full audit trail, zero invisible actions.
Tech · AsyncLocalStorage + SSE event stream · 5-minute timeout · server-side logging
03Pillar
Code and data stay yours
Spaceship installs on-premise via Docker Compose with a single command. Supabase self-hostable for data and auth. The LLM router is open: pick Anthropic, OpenAI, Google or local models via Ollama. No SaaS lock-in, no data leaving your infrastructure unless you want it to.
Tech · Tauri v2 (Rust + React) · Mastra model router · 40+ LLM providers · Docker Compose
How we build it together
05 / 08From first call to go-live, five stages.
Studio Futuro doesn't sell a packaged SaaS. Every Spaceship install is adapted to your accounting software, your flows, your people. Here's how we work.
Fase / 01
2 weeks
Operational audit
Two weeks observing processes with your staff. We identify where time is lost, where AI makes sense and where it doesn't. Output: report with priorities and ROI estimate.
Fase / 02
1-2 weeks
Module configuration
We activate the modules with highest ROI in your context. We adapt agent personas to your language, your roles, your approval policies.
Fase / 03
2-3 weeks
Tool integration
We connect Spaceship to your ERP, CRM, drive, bank. We write missing MCP connectors. We test on real data in a staging environment.
Fase / 04
1 week
Team training
Hands-on sessions with the staff who will use Spaceship. No slides — real cases on their own desktop. We also train your IT for basic maintenance.
Fase / 05
4 weeks + handover
Go-live and monitoring
Gradual rollout by department. Active monitoring for the first 4 weeks. Iteration on flows that emerge from real use. Then handover.
FAQ · Common questions
06 / 08The questions we always get asked.
Wherever you decide. Spaceship installs on-premise via Docker Compose on your own servers, or on private cloud. Supabase is self-hostable. If you pick a cloud LLM provider (Anthropic, OpenAI), only the chat messages go through them — and you can pick local models via Ollama so nothing leaves.
Two line items: one-off setup (audit + configuration + integration, scoped per project) and annual licensing per organization. No per-seat fees. LLM provider costs sit on your own accounts, under your direct control.
Yes, it's the primary use case. Docker Compose one-liner, Supabase self-hosted for auth and data, LLM router with local models via Ollama. The only thing that stays with us is the module source code — and that's released to the client at end of project.
Depends on the software. We have MCP connectors ready for the most common (Fatture in Cloud, TeamSystem, Dynamics 365). For others, we write the connector during the integration phase. If the software has documented APIs, it works.
From contract signature to first module in production: 6-10 weeks. Audit (2) + configuration (1-2) + integration (2-3) + training (1) + go-live (4 with monitoring). Depends on how many modules we activate and the complexity of your stack.
No. That's the whole point. Modules arrive pre-trained, staff write in natural language like to a colleague. 'Log invoice X', 'send me last month's report', 'find Rossi's contract'. No tricks, no syntax.
For read operations, the operator sees the answer and judges it as they would a colleague's. For write operations (categorize, send, modify), there's always an explicit human approval first. Full audit trail to reconstruct what was done and why.
Contact · Next steps
07 / 08Book a 30-minute call.
Tell us what your non-technical team does today and where time gets lost. We'll tell you if Spaceship makes sense, which modules would start first, and a rough estimate of time and cost. No sales pitch.
hello@studiofuturo.ai · Reply within 24 working hours