Real-Time Financial Insights for Uncertain Times: My Cloud Collaboration Setup with Beancount

Hey everyone,

If there’s one thing I’ve learned running my CPA practice in 2026, it’s that “uncertainty is the new certainty.” Between tariff policy shifting daily, persistent inflation pressures, and clients juggling five different economic scenarios simultaneously, the financial advisory landscape has fundamentally changed. And you know what my clients ask me most? “Can I see my numbers right now?”

The Real-Time Dilemma

Three months ago, a small business client called me in a panic. New tariffs had just been announced, and they needed to model three scenarios immediately: best case (tariffs rolled back), base case (current rates hold), and worst case (tariffs expand to more categories). They wanted live visibility into their cash position, runway calculations, and expense forecasts—not a PDF report emailed two days later.

Traditional accounting software couldn’t give them what they needed: real-time collaboration where we’re both looking at the same data, full transparency into every transaction and assumption, and version control so we could roll back scenarios and compare side-by-side. That’s when I realized I needed to rethink my entire workflow.

My Solution: Git + Fava + Cloud Deployment

After months of iteration, here’s the setup that’s transformed how I work with clients:

1. Git Repository as Source of Truth

Each client gets a private GitHub repository (GitLab or self-hosted Git work too). Their complete Beancount ledger lives there, with full version history. Every change is tracked: who made it, when, and why (commit messages are gold for audit trails).

2. Cloud-Hosted Fava Dashboard

I deploy Fava on a DigitalOcean droplet ($12/month covers several clients). Clients access their dashboard via HTTPS—no software installation, works on phones and tablets. They see their financial position updated in real-time as I push changes.

3. Collaborative Workflow

  • Morning: Client emails receipts and questions
  • Afternoon: I pull their ledger, make updates locally, test balance assertions
  • Evening: Push changes to Git, Fava auto-updates, client sees results immediately
  • Quarterly reviews: We’re both looking at the same Fava dashboard during video calls

4. Scenario Modeling That Actually Works

This is where it gets powerful. For that tariff-worried client, I created three Git branches: best-case, base-case, and worst-case. Each branch has different expense assumptions and revenue projections. I can switch between scenarios in seconds and generate comparison reports. The client logs into Fava, sees all three scenarios in the branch selector, and can explore each one independently.

What This Delivers (That Traditional Tools Don’t)

Real-time visibility without sacrificing control: Clients see live dashboards but never lose ownership of their data. Everything’s in plain text, in their Git repo.

Audit trail you can actually follow: Git history shows every journal entry addition, every correction, every scenario adjustment. When the IRS asks “how did you calculate this?”, we have receipts—literally, commit-by-commit.

Flexible enough for 2026’s chaos: One client needed to model five different revenue streams (grants, earned income, partnerships, recurring donors, one-time campaigns). Plain text + Git branches made this manageable. Try doing that in QuickBooks without creating five separate company files.

Collaborative without the cloud accounting lock-in: We’re using cloud infrastructure (for Fava) but the data isn’t locked in a proprietary SaaS. If my client wants to switch accountants, they clone the Git repo and they’re done. No export/import drama, no data loss, no vendor negotiations.

The Technical Setup (For Those Who Want Details)

  • Git hosting: GitHub private repos (free for small teams) or self-hosted Gitea
  • Fava deployment: Ubuntu 22.04 droplet, systemd service for Fava, nginx reverse proxy with Let’s Encrypt SSL
  • Security: SSH keys for Git access, HTTP basic auth for Fava (can upgrade to OAuth if needed), fail2ban for brute-force protection
  • Backup: Git gives us version history; I also run nightly borg backups to Backblaze B2 ($0.005/GB/month)
  • Client access: Most clients just use the Fava web interface. Tech-savvy ones clone the repo and use their own tools.

Real-World Impact

That tariff-worried client? We modeled their scenarios in real-time during our call. They decided to prepay suppliers (locking in current prices) after seeing the worst-case runway drop to 8 months. That decision saved them an estimated $40K based on the tariff increases that did eventually happen.

Another client—a nonprofit facing federal funding cuts—uses our shared setup to show their board real-time restricted vs. unrestricted fund balances. Board members log into Fava before meetings, come prepared with actual questions instead of asking “so, uh, how’s our cash position?”

The Learning Curve Trade-Off

I won’t pretend this is plug-and-play. Initial setup took me a weekend, and clients need basic training (“this is your dashboard, here’s how to filter transactions”). But compared to the years I spent fighting with QuickBooks permissions and Xero’s quirks, this is sustainable. And once it’s running, updates are just git push away.

Questions for the Community

I’m curious how others are handling the 2026 reality of clients demanding real-time insights:

  1. Are you using cloud-hosted Beancount setups, or is everyone still local-only?
  2. For those doing client work, how do you structure multi-client deployments? Separate repos per client? Separate Fava instances?
  3. Anyone else using Git branches for scenario modeling, or am I overthinking this?
  4. What’s your approach to client training? Do they learn Git, or do you shield them from it entirely?

The economics of 2026 aren’t getting more predictable anytime soon. But at least with the right tools, we can help our clients navigate the uncertainty with confidence—and real data.

— Alice

Alice, this is BRILLIANT! :rocket:

I’ve been doing something similar for my personal FIRE tracking, but you’ve taken it to the professional client-service level in a way I hadn’t fully considered.

My Setup: Self-Hosted on a Home Server

Instead of a cloud droplet, I’m running Fava on a home server (old desktop with Ubuntu) behind Tailscale for secure remote access. Cost: $0/month (electricity negligible). I can access my dashboard from anywhere—on my phone while traveling, from work during lunch breaks—and it’s all my hardware, my data, zero cloud dependencies.

Real-Time Portfolio Tracking in 2026’s Volatile Markets

Your scenario modeling resonates deeply. With 2026’s market volatility, I have three branches in my personal ledger:

  • optimistic: 8% annual returns, retire at 42
  • base-case: 5% returns (more realistic), retire at 45
  • pessimistic: 2% returns + recession, retire at 48+

I check all three weekly. When the market dips (like it did in February), I switch to pessimistic view, see that I’m still on track, and it prevents panic selling. This psychological benefit alone is worth the setup effort.

Dashboard Setup Details

For anyone curious, my Fava dashboard includes:

  • fava_investor plugin: Shows asset allocation, performance attribution, lots vs individual holdings
  • Custom BQL queries: Net worth by month (graphed over 5 years), savings rate trending, “years to FI” countdown
  • Automated imports: Plaid2text for bank accounts, custom scripts for Vanguard CSV downloads
  • All tracked in Git with daily automated commits

When my spouse asks “are we on track for early retirement?”, I pull up the dashboard on my phone and we’re both looking at real data, not guesses.

Why This is the Future

Traditional financial advisors charge 1% AUM to show you quarterly PDFs that are stale the moment they’re generated. This approach—real-time, transparent, version-controlled—represents the future of transparent financial management.

You own your data. You see every transaction. You can model scenarios instantly. And if you’re technical (or working with someone technical like Alice), you can build exactly the analytics you need, not what some SaaS vendor decided to productize.

Alice, two questions for you:

  1. Do your clients ever want to make direct edits to their ledger (beyond using Fava’s UI), or do you handle 100% of the transaction entry?
  2. How do you handle reconciliation when clients have questions? Do you hop on a call and screenshare Fava, or do they learn enough to navigate themselves?

This is exactly the kind of practical innovation our community needs more of. Thanks for sharing! :raising_hands: