Autonomous modernization for regulated legacy shops

Modernize the .NET Framework systems nobody dares touch.

Forgestus takes your classic .NET monolith to modern .NET — slice by slice, every slice proven behavior-identical by pinned tests, every action in an immutable audit ledger, all of it inside your network on models you control.

Behavior-parity gates · Windows Fargate runners Your network, your models · HIPAA/HITRUST enforced at runtime
Sound familiar?

Three problems every AI coding demo skips.

If your stack is older than the tools, your auditors ask harder questions than your engineers, and your roadmap outruns your headcount — you're who we built this for.

Your stack outlived the tooling.

The .NET Framework monolith pays the bills, and nobody wants to be hired into it. Every AI coding tool you've evaluated is JavaScript-and-Python-first — point it at MSBuild, NuGet, or a .sqlproj and the demo ends. Meanwhile the legacy backlog only grows.

THE FORGESTUS ANSWER → Forgestus migrates classic .NET Framework to modern .NET slice by slice on a Windows Fargate runner — MSBuild, NuGet, vstest, WCF, WinForms, SSDT and all. And every migrated slice ships with pinned characterization tests proving the old behavior still holds on the new runtime. Modernization you can prove, not just trust.

Compliance keeps killing the AI pilot.

You want the velocity; your auditors want to know who wrote each change, what it touched, and what it cost. "An agent did it" is not an answer a regulator accepts — so every AI initiative stalls at the governance review, and opaque token bills finish it off with the CFO.

THE FORGESTUS ANSWER → A permanent audit ledger records every model call, every retry, every dollar — exportable to CSV (SIEM connectors on the roadmap). GUARDRAILS.md constraints and HIPAA/HITRUST-aware coding rules are enforced at runtime and hard-flagged by reviewers — not bolted on after. Velocity and governance aren't a tradeoff here. Autonomy your compliance team can read.

The backlog grows faster than the team.

Roadmap items sit for quarters because the two engineers who can safely touch the core are always on fires. Hiring takes six months in your vertical — if the candidate exists at all — and contractors won't learn the domain before the contract ends.

THE FORGESTUS ANSWER → File the epic as a GitHub issue and type /proceed. Planner, architect, backend, web, and mobile agents build the stories in parallel, Opus reviews every change, and the epic lands deployed in your test environment — while your senior engineers stay on the work only they can do.

How modernization works

From assessment to cutover, autonomously.

No big-bang rewrite. Forgestus modernizes your monolith slice by slice — and proves each slice behaves identically before it ships. Every step durable, every retry idempotent, every cost in the ledger.

Assess

A read-only scan maps your architecture and business rules, ranks risk by churn × complexity × coverage, and produces a dependency-ordered strangler plan with difficulty scores. Nothing is written to your code.

Pin the baseline

Before touching anything, Forgestus runs your existing tests on the old .NET Framework runtime and pins the result — the behavior every migrated slice will be measured against.

Migrate a slice

Agents port the slice to modern .NET on a Windows runner — applying recipes like CoreWCF where they fit, and flagging anything non-portable for a human instead of guessing.

Prove parity & cut over

A green parity report — new runtime vs the pinned baseline — gates the merge. The cutover board moves the slice monolith → shadow → live. Repeat until the monolith is modern.

Every slice lands with a green parity report as its receipt — start with an assessment →

Beyond migration

Once it's modern, keep shipping.

The same deterministic agents that migrated your monolith deliver your everyday epics — across 11 stacks, with the same audit ledger and Opus review gates. File a GitHub issue, type /proceed, and the pipeline ships it.

File the epic

Open a GitHub issue. Describe the work in plain language. Forgestus's planner agent breaks it into stories, posts a plan, and waits for your /proceed.

Build in parallel

Architect, backend, web, and mobile agents work on isolated story branches at the same time — each with full repo context, each running the right toolchain for the right stack.

Review with Opus

Reviewer agents audit every PR. halt verdicts trigger auto-retry of the implementing agent. advisory verdicts get logged. approve moves the epic forward.

Deploy and validate

Forgestus merges to develop, runs your deploy workflow, smoke-tests the deployed routes, and advances to the test environment on green. Failures route to devops-fix with the actual error in scope.

Why Forgestus

Six things procurement asks for. We check all six.

11 production stacks

Node, modern .NET, classic .NET Framework on Windows Fargate, React, Next.js, Angular, Blazor, Flutter, Unity, WinForms/WPF, SSDT. The enterprise long tail, first-class.

Deterministic state machine

Every transition durable in DynamoDB. Every retry idempotent. When an agent crashes, the orchestrator recovers without losing work. Crash-safe by construction.

Permanent audit ledger

Every model call, every retry, every dollar — written to AUDIT# rows that never expire. CSV export, regulator-ready (SIEM connectors on the roadmap).

Self-healing recovery

Stuck epics auto-diagnose and fix themselves where safe. When an agent hits a target that doesn't exist, it pauses and asks instead of burning its budget — you /answer and it resumes. Every blocker surfaces a one-glance diagnosis with the exact fix command. No "wedged forever" mode, no log spelunking.

HIPAA + HITRUST as config

Add compliance: [hipaa] to forge.yaml — every agent inherits the coding controls. Enforced by the runtime, audited by the reviewers.

Bring your own key

Anthropic, AWS Bedrock, OpenAI, or Vertex — your contract, your rate, your data-handling. We run the infrastructure; the flat platform fee covers it. Model tokens are the only usage cost, billed to you by your vendor — no bundled credits, no markups.

Stack reach

From Blazor to Unity. One platform.

Classic .NET Framework 4.x on a Windows Fargate runner. Lambda dotnet8 on Linux. Flutter, Unity, Angular, Next.js — every stack with a first-class agent prompt and a deterministic QA runner.

Node.jsjest · tsx
.NET 8xunit · sdk-style
.NET Frameworkmsbuild · vstest
Reactjest · vitest
Next.jsapp-router · ssr
Angularkarma · jasmine
Blazorbunit
Flutterflutter test
Unityeditmode · playmode
WPF / WinFormsxaml · winui
Windows Servicebackground
SSDTtsqlt · dacpac · LocalDB
C# CDK.NET infra-as-code
MCP serversaws-docs · cloudwatch
Custom agentsyour prompts · your steps
Multi-modelAnthropic · Bedrock · OpenAI · Vertex · Meta
Pricing

A flat platform fee. Infra included. Bring your own key.

Three tiers, repo-gated, with epics included in every plan. We run the infrastructure — the platform fee covers it; the only usage cost you pay is your own model tokens (bring your own Anthropic key, or AWS Bedrock / OpenAI / Vertex on Enterprise). Annual billing 15% off; two-year price-lock 25% off.

Team

$1,000/month
For a small product team with backend, web, and mobile repos.
  • Up to 3 connected repos · single GitHub org
  • 20 epics / month included · $40 per additional epic
  • Core stacks: Node, React, Flutter
  • Dev deploy automation · 90-day audit retention
  • GUARDRAILS.md enforcement · Opus reviewers
  • BYOK: your Anthropic key — model tokens billed to you directly
  • Email support · 1 business day response
Start with a pilot
Need more repos or stacks? Upgrade to Business →

Enterprise

Talk to us
For regulated, polyglot, and legacy-stack engineering organizations.
  • Unlimited repos · unlimited orgs · unlimited epics
  • + Unity (EditMode / PlayMode) · custom stack onboarding (Rust, Go, Elixir…)
  • VPC deployment in your AWS — your infra, your data boundary
  • Model fallback chains for rate-limit resilience
  • Multi-model: Anthropic + Bedrock + OpenAI + Vertex + Meta
  • Advisor tool: fast agents call in Opus on the hard calls
  • Custom agent prompts · custom pipeline steps (early access)
  • Auto-refreshed epic learnings from your repos
  • CSV audit export (Splunk / Datadog / S3 connectors on the roadmap)
  • Direct-with-founder support · priority response
Talk to sales
managed compute available as add-on
FAQ

Questions enterprise procurement actually asks.

What stacks do you actually support?
Node, modern .NET 8, classic .NET Framework 4.x on a Windows Fargate runner, React, React + Vite, Next.js, Angular, Blazor, Flutter, Unity, C# Windows desktop (WPF / WinForms), C# Windows Services, and SSDT. The full .NET family — including classic .NET Framework on Windows runners — is available on the Business plan; Unity and custom stack onboarding (Rust, Go, Elixir, Kotlin/Android-native) are Enterprise.
What counts as an epic, and what happens at my monthly limit?
An epic is one GitHub issue you file and /proceed — however many stories the planner breaks it into. Team includes 20 epics/month, Business 75; beyond that, additional epics are billed per-epic ($40 Team, $30 Business) — nothing stops mid-flight, and in-progress epics always run to completion. Epics on Windows runners (classic .NET Framework, WPF/WinForms, SSDT) count as two, reflecting the heavier compute. Enterprise plans are unlimited because they run in your own AWS account.
Do I really bring my own AI vendor key?
Yes, on every tier. Forgestus runs against your Anthropic key, AWS Bedrock credentials (Business+), OpenAI key (Enterprise), or Google Vertex credentials (Enterprise). Your AI vendor invoices you directly at your contract rate — model tokens are the only usage cost you pay. On Team and Business we run everything in our managed cloud and the flat platform fee covers all infrastructure (Fargate runners, orchestration, storage) — you never see an AWS bill. Only Enterprise customers who self-host in their own AWS account pay AWS directly, and they can opt into Forgestus-managed compute with a 10–15% pass-through for a single invoice.
Does Forgestus support HIPAA / HITRUST?
Yes. Add compliance: [hipaa, hitrust] to forge.yaml and every agent inherits the coding constraints — PHI logging rules, KMS encryption, audit-trail patterns, least-privilege IAM, DLQ requirements. Enforced by the runtime, hard-flagged by the reviewer agents. SOC 2 and FedRAMP are on the roadmap; PCI-DSS available as custom stack onboarding.
What happens when an agent gets stuck?
It doesn't wander or burn its budget. If an agent can't find a target the story names — a screen, form, or endpoint that doesn't exist in the repo — it pauses, asks the question on the issue, and waits. You reply with /answer and it resumes from where it stopped. For everything else, every blocker carries a structured diagnosis right in the portal: what's wrong, why, and the exact operator command to fix it (/retry, /restart, /proceed, /unblock). No "wedged forever" epics, no CloudWatch digging.
Can I customize the agent pipeline?
On Enterprise, yes. You can prepend project-specific constraints to any built-in agent's prompt — or override it entirely — without losing the battle-tested defaults, and Forgestus auto-refreshes a per-repo learnings skill weekly, so the agents get smarter about your codebase over time. Fully custom agent types injected at specific pipeline stages (e.g. a security pass after backend) are in early access — ask us about it.
What is the advisor tool?
A way to get Opus-grade judgment at fast-model cost. A quick executor agent (Haiku or Sonnet) runs the bulk of the work, but when it hits a genuinely hard decision it consults a higher-intelligence Opus advisor mid-task — and keeps going with the answer. You pay the advisor rate only for the moments that need it, not for the whole run. Enable it per-project and per-agent in forge.yaml; advisor token usage and cost are itemized separately on every audit row. Available on Enterprise, on your Anthropic key or Claude on AWS.
Is this self-hosted? Can I run it in my own VPC?
Enterprise customers can deploy Forgestus in their own AWS account as a VPC deployment — the orchestrator Lambda, DynamoDB, ECS Fargate, and S3 storage all run in your account, with Forgestus providing the agent images and release cadence. Team and Business tiers run in our managed cloud.
How do I get started?
Fill out the form below — the One Epic pilot is two weeks, one real epic from your backlog delivered end-to-end on your GitHub repo. We'll respond within one business day with a scoped pilot proposal. Pricing credits against your first three months of platform fee on conversion.
The One Epic pilot

Run one of your real epics in two weeks.

Two weeks. One real epic from your backlog, delivered end-to-end on your GitHub repo. Includes forge.yaml authorship, two operator training sessions, and a written cost + outcome report. Tell us about your stack — we'll get back to you within one business day with a scoped pilot proposal.

One or two sentences is enough. We'll scope it together on the intro call.

By submitting, you agree to be contacted about your pilot. No spam, no list-sharing.

Got it — talk soon.

We'll review your epic brief and reply within one business day from contact@forgestus.ai with a scoped pilot proposal.

Forgestus runs on its own compute during the pilot — no AWS or Anthropic setup required to evaluate.  ·  Just have a question? Email us directly.

"You can outsource your thinking — but you can't outsource your understanding." Forgestus automates the work. Your team keeps the judgment.