OpenFrame v0.7.8 is live. Here's what shipped, what it means for your stack, and where we're headed.

For those catching up: OpenFrame is an AI-native MSP platform that replaces the patchwork of vendor tools most MSPs run – RMM, ticketing, monitoring, remote access, MDM – with one unified system. Self-host it for free or run it managed. v0.7.8 builds on that foundation with a tighter API layer, smarter AI agents, more deployment flexibility, and the developer experience that's attracting over 1,800 MSPs to the platform platform.

One API Instead of Eight

The biggest architectural win in v0.7.8 is the unified API. Instead of juggling separate interfaces for every tool in your stack, you get one consistent API across RMM, ticketing, monitoring, and remote access.

For MSPs migrating off ConnectWise or Kaseya, this changes the math on integration work. You're not swapping five vendor APIs for five open-source APIs – you're replacing them with one. OpenFrame handles the orchestration underneath, translating your requests into the right calls to osquery for endpoint visibility, Zabbix for monitoring, and MeshCentral for remote access.

The API follows OpenAPI 3.0 specs, so generating client libraries in whatever language your team uses is straightforward. Build custom workflows, pull device inventory into your own dashboards, or automate ticket routing – all through the same endpoints with consistent auth and error handling.

AI Agents: Fae and Mingo

v0.7.8 includes two AI agents that split the work between client-facing and backend operations.

Fae handles intake. When an end user reports an issue, Fae processes the natural language input, pulls context from the system – logs, screenshots, device info – and creates a structured ticket. That cuts the back-and-forth your techs normally spend on triage.

Mingo handles execution. It generates scripts in PowerShell, Bash, or Python based on the task, using retrieval-augmented generation trained on common MSP remediation patterns. Every generated script gets logged and can be reviewed before it runs. No black boxes.

Honest assessment: the AI agents are strong on L1 and L2 tasks – password resets, disk cleanup, basic troubleshooting, patch management. Complex scenarios still need a human in the loop. We're building Fae and Mingo as force multipliers for your techs, not replacements. The agents can run locally or connect to external LLM providers, depending on your security requirements and budget.

Deployment: Your Call

Three options, same platform:

Self-hosted (free): Deploy the full platform on your infrastructure using Kubernetes or Docker Compose. Helm charts and infrastructure-as-code templates are in the GitHub repo. You get complete control over data residency, customization, and scaling – which matters if you're dealing with compliance requirements.

Managed: Flamingo runs it for you with per-device pricing. Automatic updates, monitoring, and support included. You still get full API access and the ability to extend functionality.

Hybrid: Run sensitive data and core services on-prem while offloading components like AI agents to Flamingo's managed environment. Useful for organizations navigating regulatory gray areas.

The platform uses a microservices architecture, so each component – RMM, ticketing, monitoring, AI agents – runs independently. Scale what you need, leave the rest alone.

What Ships in the Core

OpenFrame v0.7.8 includes modules that would normally require separate vendor subscriptions:

RMM – Built on osquery for cross-platform endpoint visibility. Real-time device inventory, patch management, and remote command execution across Windows, macOS, and Linux. The query interface lets you write SQL-like queries against your entire fleet.

Ticketing and PSA – Purpose-built for multi-tenant MSP environments with SLA tracking, time tracking, client portal, and billing integration hooks. Not a generic ticketing system bolted onto an MSP workflow.

Monitoring and alerting – Prometheus and Grafana integration with pre-built dashboards for common MSP scenarios. The alerting engine ties into the ticketing system for automatic ticket creation.

Remote access – Browser-based connections through Apache Guacamole with session recording and audit logging. No client software needed.

MDM – iOS and Android device management, including app deployment and configuration profiles.

SIEM – Security event correlation pulling data from endpoints, network devices, and cloud services into one view.

Each module plugs into the unified API, so you get both granular control per tool and system-wide consistency.

What's Next

OpenFrame Generations Roadmap

Here's where OpenFrame is headed:

Multi-platform integration – The adapter architecture we built for FleetDM integration extends to other device management backends. Same declarative policy interface, more platforms underneath.

Policy composition – We're expanding the policy language to support inheritance, composition, and templating. Reusable components that can be composed into complex configurations without duplication.

Community extensions – We're opening extension points for community-contributed policy libraries, custom reconciliation logic, and integration adapters. The architecture supports this without compromising core stability.

These aren't vague roadmap slides. They're direct extensions of the infrastructure decisions already made in v0.7.8.

OpenFrame is free to self-host. Spin it up from GitHub, break it, deploy it. If you want the managed version or have questions, the OpenMSP Slack community has 200+ MSPs sharing implementation patterns and troubleshooting in real time.

Frequently Asked Questions

Q: Is OpenFrame truly open source, or is it open-core with paid features?
A: The core platform is fully open source under a permissive license. The self-hosted version includes all core functionality at no cost. The managed service and AI agents are where Flamingo monetizes, but you can run everything yourself if you prefer.

Q: What's the learning curve for developers new to the platform?
A: If you're familiar with modern cloud-native architectures (Kubernetes, microservices, REST APIs), you'll be productive within a few days. The CLI and documentation are excellent, and the unified API significantly reduces complexity compared to traditional MSP stacks.

Q: How does OpenFrame compare to building a custom solution from scratch?
A: OpenFrame gives you a pre-integrated stack of best-in-class open-source tools with a unified API layer. Building this yourself would take months and require deep expertise in each component. The platform handles the integration complexity, letting you focus on customization and business logic.

Q: What's the community like? Is this a one-person project or a real ecosystem?
A: Flamingo has backing from Focal VC and Array VC ($2.2M in pre-seed funding), with over 1,800 MSPs on the waitlist and 150+ active beta users. The Slack community is active (500+ memebers), and the team is responsive. This isn't a side project-it's a well-funded startup with serious momentum.

Q: Can I contribute to the project? How does governance work?
A: Yes, contributions are welcome. The project uses standard GitHub workflows (pull requests, issues, discussions). While Flamingo maintains the core platform, they're actively encouraging community contributions for extensions, integrations, and documentation.

Q: What about vendor lock-in? If Flamingo disappears, what happens?
A: Because the core is open source and uses standard open-source components underneath, you're not locked in. You could fork the project and maintain it yourself if needed. The architecture is transparent, and all data is stored in standard formats (PostgreSQL, InfluxDB, etc.).

Q: How mature is the AI agent functionality? Is it production-ready?
A: The AI agents are in active development and improving rapidly. Early adopters report good results for common tasks (password resets, disk cleanup, basic troubleshooting), but complex scenarios still require human oversight. Treat the AI as an assistant, not a replacement for skilled technicians-at least for now.

Q: What are the hardware requirements for self-hosting?
A: Minimum viable deployment: 3 nodes with 4 CPU cores and 16GB RAM each. For production with 500+ endpoints, plan for 5-7 nodes with 8 cores and 32GB RAM. Storage requirements depend on retention policies for logs and monitoring data-budget 1-2TB for a typical deployment.

Conclusion

OpenFrame v0.7.8 represents a significant evolution in MSP platform architecture-one that prioritizes developer experience, deployment flexibility, and economic sustainability. By combining open-source infrastructure with modern AI capabilities and a unified API layer, the platform addresses real pain points that have plagued the industry for years.

For technical teams evaluating MSP platforms, OpenFrame offers a compelling alternative to legacy vendors: full control over your stack, transparent pricing (or no pricing for self-hosted), and the ability to customize every aspect of the platform. The architecture is sound, the performance is solid, and the community momentum is real.

The platform isn't perfect-it's still relatively young, and some features are more mature than others. But the trajectory is clear, and the team is shipping improvements rapidly. If you're tired of vendor lock-in, frustrated by integration complexity, or simply curious about how modern MSP platforms should be built, OpenFrame v0.7.8 deserves serious evaluation. Explore the Flamingo knowledge base for implementation guides and technical documentation.


Kristina Shkriabina

Kristina Shkriabina

Kristina runs content, SEO, and community at Flamingo and OpenMSP. She spent years as a correspondent for Ukraine's Public Broadcasting Company before making the jump to tech. Now she covers MSP stack decisions and strategy. You can connect with her in the OpenMSP community or on LinkedIn.