A GitHub alternative for knowledge work
Palette gives your team a shared place for documents, instructions and decisions. Work with your AI, review changes and save useful work for everyone to build on.
If you’re keeping company knowledge in GitHub, the reasoning makes sense. Your team needs a common place to work from, a way to review changes and a record of how things developed.
We’ve borrowed those ideas and made them easier to use across the company.
Why teams use GitHub for knowledge work
A brief, a process or a set of instructions can live in a plain text file, such as Markdown. People can read it, agents can work with it, and changes can be compared and reviewed.
Clear filenames and folders help explain what each document is for. The same material can be used by different AI tools, so the knowledge your team builds remains useful when you change tools.
GitHub combines those files with a way to propose changes, review them and keep a shared version. Its own team uses GitHub flow for documentation, policies and its roadmap.
An agent can read the agreed positioning before drafting a proposal, follow written instructions and contribute changes for review. Useful work goes back into the files, where colleagues and their agents can use it.
Keeping it current takes the whole team
GitHub can work well when everyone is comfortable with its workflow. Across a wider company, that familiarity varies.
The person who knows the client may never have opened a pull request. The person responsible for the sales playbook may need help getting an update into the repository. Their knowledge reaches the shared files only when someone else puts it there.
Over time, maintaining the source of truth can become one person’s job.
Files alone don’t keep knowledge current. Someone still needs to own the guidance, update it when things change and review what belongs there. The people closest to the work need a straightforward way to contribute.
The ideas we borrowed from GitHub
Palette brings shared files, separate working sessions, review and history to everyday documents and AI work.
| The idea in GitHub | How it works in Palette |
|---|---|
| A shared repository | A shared space for a team, client or project |
| A branch for working on changes | A session where you can draft and experiment separately |
| Reviewing proposed changes | Seeing what was added, edited or removed before saving |
| Merging agreed work | Saving changes back to the shared space |
| A history of contributions | Activity showing what changed, who contributed and why |
You can try a different approach, rewrite a plan or have an agent prepare several drafts in a session. When the work is ready, review the changes and save what belongs in the team’s space.
Drafting and experimenting remain part of the work. They don’t have to change the files everyone else relies on.
From updated positioning to the next proposal
Say marketing updates how your company describes its product.
They work on the wording, review the changes and save the agreed positioning. A salesperson then uses that positioning, the client brief and the pricing guidance to prepare a proposal with their AI.
The salesperson reviews the proposal and saves it in the client’s space. If the work reveals a useful clarification, that can be added to the guidance too.
The next proposal can start from those files. Nobody needs to find the original chat or ask marketing to explain the change again.
Work from the AI tools you already use
Connect Claude, ChatGPT, Gemini or Mistral to your shared spaces, or work in the Palette app.
Your colleagues can use different AI tools while working from the same documents and instructions. The team’s knowledge stays in the space as people change tools.
Instructions explain how to work: where a proposal belongs, which pricing guidance to use and how to write in the company’s voice. As you work, agents can help update existing documents and record useful corrections.
You choose the work worth keeping. A correction saved in the space can guide the next task, including one started by a colleague.
Choosing between GitHub and Palette
If your team already maintains its knowledge comfortably in GitHub, that can be a good setup.
Palette is worth considering when you want more people to contribute directly, particularly colleagues who work with documents and AI but don’t use GitHub.
You can also use both: GitHub for your software projects, Palette for the briefs, research, playbooks and decisions shared across the company.
Start with one small space
Choose a project or recurring piece of work. Add a few relevant documents and instructions, connect your AI and try a task you already need to do.
Review the result, save what’s useful and invite a colleague to build on it.