Regen v1.0 roadmap — migration tools, MCP server, and the path to GA

[singret]

What we are building toward v1.0

We shipped v0.11.0 last week. Here is what we are working toward for v1.0 GA, currently targeting late July 2026.

This is our public OSS roadmap — sharing it here so the community can see what is coming, flag what matters most, and help us test as things land.

---

The opportunity we are building for

Grafana OnCall is being archived in March 2026. Around 50,000 self-hosted teams need a new home. At the same time, PagerDuty and Opsgenie cost $30–50/user/month — $120k/year for a 200-person engineering org. The case for a self-hosted, fully featured, permanently free alternative has never been stronger.

Regen already covers the full lifecycle: alert ingestion (Prometheus, Grafana, CloudWatch, generic webhooks) → on-call schedules + escalations → incident coordination → Slack/Teams sync → AI summaries + post-mortems. v1.0 is about making all of that polished enough that any team can adopt it in under 15 minutes.

---

What is landing in v1.0

1-click migration from PagerDuty and Opsgenie

The biggest blocker for adoption is migration friction. We are building CLI-based importers that pull your schedules, escalation policies, and service mappings in one command:

regen import pagerduty --api-key=<key> --dry-run
regen import opsgenie --api-key=<key> --dry-run

Dry-run shows exactly what would be created before writing anything. Idempotent so you can re-run safely.

If you have migrated from PagerDuty or Opsgenie before, we would love your input on what the importer needs to handle. Edge cases in schedule layers, escalation timeouts, service mappings — drop them below.

Onboarding wizard (zero to first incident in under 15 minutes)

A new self-hoster should be able to go from docker compose up to seeing their first real incident — Slack channel created, on-call engineer notified — in under 15 minutes without reading the docs. We are building a guided wizard: connect Slack → import or create a schedule → send a test alert → confirm incident appears.

Mobile-responsive UI + PWA push notifications

On-call engineers get paged at 3am. The current UI is not great on mobile. We are fixing that with proper Tailwind responsive breakpoints and adding web push notifications (PWA) so your phone shows the alert without needing the tab open — no app install required, works on iOS and Android.

More AI provider options (BYO key)

We currently support OpenAI for incident summaries and post-mortem drafts. We are adding:

• Anthropic Claude — ANTHROPIC_API_KEY
• Ollama — fully local, no data leaves your infrastructure. For regulated industries where cloud AI is not an option.

All providers use the same interface. Your incidents stay in your infrastructure regardless of which you pick.

Schedule quality-of-life improvements

• Override UI polish — the current override flow requires too many steps. We are rebuilding the creation form and calendar visualisation so it is usable without docs.
• Holidays in schedules — import public holidays via ICS feed per country. The rotation engine automatically skips engineers on their local public holidays.

Internal status page

Every tenant gets a public status URL (https://<your-host>/status/<slug>) that lists active and recently resolved incidents. No configuration required, no customisation in v1 — just a shareable link you can paste in your internal comms.

Documentation that is actually tested

We are committing to documentation that is run on a clean install — not just written. This means:

• Fresh deploy guide (Docker Compose + Kubernetes/Helm)
• Migration guides: PagerDuty → Regen, Opsgenie → Regen, Grafana OnCall → Regen
• Failure runbooks for the top 10 operational failure modes

MCP server (agent-native)

This is the longer-term architectural bet. Regen will expose a read-only MCP server so external AI agents (Claude, GPT, custom bots) can query your incident history, triage context, on-call status, and service health profiles natively:

search_incident_history(service, time_from, time_to, severity, query)
get_triage_context(incident_id)   → timeline + alerts + on-call + related incidents
get_current_oncall(service)       → current person + next handoff
get_service_health_profile(service) → MTTA/MTTR + top failure modes

Bearer auth, per-tenant scoping, schema validation. This lands in W11–W12 (early July).

---

Timeline

Week	Dates	What ships
W1–W2	Apr 27 – May 9	Foundation + schedule overrides + mobile UI
W3	May 11 – 16	PWA push + onboarding wizard + AI providers
W4–W5	May 18 – 30	PagerDuty import + custom fields + status page
W5–W6	May 25 – Jun 6	Opsgenie import + Slack improvements + docs
W11–W12	Jul 6 – 18	MCP server + 4 agent tools
W13	Jul 20 – 26	v1.0 GA tag + Helm chart published

---

How you can help

If you are migrating from Grafana OnCall, PagerDuty, or Opsgenie:
Tell us what your setup looks like. Complex rotation patterns, custom escalation logic, service mappings — the more edge cases we know about before building the importers, the better they will be.

If you have tried Regen and hit friction:
Open an issue or drop a comment here. The onboarding wizard is only as good as the real blockers we know about.

If you want to test early:
PRs for the items above will be tagged needs-testing when they are ready for review. Watch the repo to get notified.

If you are interested in the MCP server:
We want to know what tools would be most useful. The four above are our starting point — what queries do your AI workflows actually need?

---

Install Regen today

git clone https://github.com/FluidifyAI/Regen.git
cd Regen
docker compose up

Or pull the latest image: docker pull ghcr.io/fluidifyai/regen:latest

Full docs: https://github.com/FluidifyAI/Regen#readme

---

Questions, feedback, or edge cases we should know about — drop them below.