Pi Network has expanded SoloHost with OpenClaw and an Atlassian MCP Server, extending Pi Desktop beyond blockchain-node operations into self-hosted AI and developer workflows. In its official announcement, Pi said both applications are now featured in SoloHost following the release of Node version 0.6.2. The update positions existing Pi hardware as a platform for running local software in addition to supporting the network itself.
SoloHost provides a containerized environment for discovering, installing and running self-hosted applications without requiring users to manually configure as much Docker or server infrastructure. Pi says the platform is available to more than 420,000 Node runners, although SoloHost remains in beta and uses an open, permissionless publishing model. The broader utility comes with an explicit warning that users remain responsible for assessing individual packages before installation.
OpenClaw Brings Local AI to Pi Desktop
OpenClaw operates as a local AI agent that can use either a model running on the user’s own hardware or external services such as ChatGPT and Claude. When cloud models are selected, requests still leave the device for processing by those providers, while OpenClaw’s memory remains stored locally. Pi is therefore offering local control over the agent environment without claiming that every AI workload stays entirely on-device.
Running OpenClaw through SoloHost also places the agent inside a container that restricts its default access to unrelated files and system resources. Users can still explicitly authorize external services or additional resources, so containment does not eliminate every security risk. The design reduces unnecessary machine-level access rather than turning the agent into a fully isolated or risk-free application.
The addition builds on Pi’s earlier support for Hermes, another local AI agent available through SoloHost. OpenClaw is positioned more around personal-assistant workflows and connections across tools and messaging services, while Hermes emphasizes persistent memory and learning from previous interactions. Together, the applications broaden Pi Desktop’s role from node software toward a more general self-hosted AI environment.
MCP Connects AI Tools With Jira
The second application targets more specialized workplace automation. Pi’s Atlassian MCP Server lets users run a local server through which MCP-compatible tools such as Cursor, Claude Desktop, Claude Code and Codex can interact with Jira. The local deployment gives teams another way to connect AI agents with project-management workflows while keeping the intermediary server on hardware they control.
The technology is based on Model Context Protocol, an open standard for connecting AI applications with external tools and data sources. Atlassian’s own MCP documentation explains that its broader Rovo MCP infrastructure can give compatible AI tools controlled access to Jira, Confluence and Compass for actions such as searching work items or creating content. Pi’s SoloHost implementation is specifically described around Jira, so support for Atlassian’s wider product suite should not be assumed from the Pi release alone.
Node 0.6.2 also introduced improvements including UPnP-based automatic port configuration and a new port-checking tool, reducing some of the setup required for locally hosted applications. The cumulative direction is clear: Pi is trying to make its desktop software useful for self-hosted AI and developer infrastructure, not solely blockchain validation. Whether those applications gain meaningful usage will depend on adoption beyond the initial availability announcement.
Natalie Pierce tracks the parts of crypto that move fast and rarely wait for everyone to catch up. From South Africa, she covers DeFi, AI crypto, hacks, airdrops, sentiment and emerging narratives, especially when user behavior and protocol risk start to overlap.
Her reporting is built for messy sectors. Natalie looks at incentives, reactions, security concerns, social momentum and early signs of traction without pretending every new trend is already proven. Her voice is clear and accessible, but careful enough for areas where excitement can outrun the facts very quickly.