Keep your Claude workflow. Connect the whole organization.
Claude Projects brings context and ongoing work together around Claude. Lit Factory connects that work to the people, coding agents, decisions, and tools across your organization—on a shared Floor you can run on your infrastructure.
Independent comparison by Lit Factory · Reviewed September 30, 2026. Claude and Claude Code are Anthropic products. Lit Factory is not affiliated with or endorsed by Anthropic.
Choose Claude Projects when…
You want a focused place to develop work with Claude, keep project context together, and use its coordinator to delegate related tasks.
The work spans several people, agent environments, and tools. You need shared ownership, reviews, operational knowledge, and a way to improve the handoffs between them.
Claude’s offering includes established Projects, the redesigned Projects beta, Cowork, and Claude Tag. The comparison names each surface so you can see how it fits alongside a shared Lit Factory Floor.
Where work comes together
Claude
Projects organizes conversations, reference material, and instructions around a body of work. The redesigned experience adds coordination across parallel Claude Code threads.
A shared Floor connects work, ownership, discussions, and decisions across your organization. People, Light, and your coding agents work from the same operational context.
The Projects coordinator delegates to Claude Code threads, with project resources shared across those threads. Local threads can connect through Remote Control.
Keep the harness you have tuned and optimized for your workflow—Claude Code, Codex, Cursor, or OpenCode—with your configured skills, tools, and team instructions. Built-in MCP connections bring these existing agent environments onto the same Floor.
Set an autonomy band on a task. Lit Factory passes behavioral instructions into the connected agent workflow, including your existing Claude Code session. Work interactively, delegate more autonomy, and step in at any point, with ownership and review connected to the task.
The Brain keeps operational knowledge alongside work and decisions. Dynamic channels organize matching work. Metrics reveal waiting and handoffs, while Optimize suggests automations with a prepared request for Light—review it and press Enter to get started.
Connect work and discussions across the Floor, Slack, GitHub, Linear, and other tools. Coordinate follow-ups and reviews across teams while keeping the relevant context with the work.
Run the shared Lit Factory platform on Kubernetes in your own cloud. Bring your database, identity provider, and model endpoints. Light can use commercial providers or self-hosted models, while external coding agents keep their own model configuration.
Develop the idea with Claude. Move the work together.
Shape the brief.
Use Claude Projects to research a problem and develop a proposal. Bring the relevant brief onto the Floor when it needs shared ownership.
Keep your harness.
Connect Claude Code through MCP. Developers continue in the environment they have optimized, while the task’s autonomy band guides their agent’s participation.
Carry it through.
Coordinate implementation, a teammate’s review, and a decision in Slack. Keep ownership and context attached to the work through completion.
How Lit Factory works
A shared workspace that helps improve how work gets done.
The Floor connects the work, its context, the people and agents responsible, and the next improvement. Start with one recurring workflow and measure what changes.
01 / Observe
See where work waits.
Follow work across connected tools. Dynamic channels organize matching work; metrics surface queues, interruptions, and review waits.
02 / Understand
Give agents a shared Brain.
Keep knowledge, decisions, and source context available to the team. Light, the built-in agent, can use that context alongside connected tools.
03 / Act
Set the responsibility.
Choose an autonomy mode for agent work. Coordinate Light, external agent clients, and recurring automations around the same work and human decisions.
04 / Optimize
Turn a bottleneck into a next step.
Optimize suggests automations with a prepared request. Review it with Light, press Enter to send it, and confirm the setup before enabling it.
Suggestions start a conversation, not an unattended change. Compare the intended autonomy with recorded participation, then adjust the workflow as evidence accumulates.
Which Claude Projects experience does this compare?
Both the established project workspace and Anthropic’s redesigned Projects coordinator. We also identify adjacent Cowork and Claude Tag capabilities where they matter to team workflows. The linked Anthropic guides describe each experience.
Can we use Claude Projects and Lit Factory together?
Yes. Use Projects for work you want to develop with Claude, then bring the relevant brief or output onto a Floor for team ownership, implementation, and review. Claude Code connects to Lit Factory through MCP, so developers can participate from their existing agent environment.
Do developers need to replace their Claude Code setup?
No. Keep the harness you have tuned for your workflow, including skills, tools, and team instructions. Lit Factory connects that environment to shared work and passes task autonomy instructions into the connected agent session. You can collaborate interactively, delegate, and step in whenever needed.
Does Claude already support collaboration and automation?
Yes. Cowork has shared projects and scheduled work, and Claude Tag supports team work in Slack. Lit Factory is relevant when you want one shared workflow across different coding agents, people, and work systems, with organization-wide context and task ownership.
How does Lit Factory pricing work for a team?
Lit Factory’s cloud pricing pools Sparks across unlimited humans and agents without a seat fee. Enterprise deployment has separate commercial terms, and external model-provider charges may apply. Your team can choose the model and deployment arrangement that fits its work.
Bring the workflow your team wants to improve.
We will map the people, agents, tools, and handoffs around it, then show how a shared Floor keeps the work moving.