A free decision tree maker should let you draw a question, label each branch, and reach an unambiguous outcome. For a file you can keep and edit locally, start with draw.io. For an existing team workshop, consider Miro or FigJam. Lucidchart fits a small diagram within its free document and shape limits; Canva fits a tree that will become part of a presentation. CodePic provides a hand-drawn whiteboard starting point, while MindMeister suits a branching outline that does not need formal flowchart notation.
This comparison is about visual decision trees for troubleshooting, policies, and planning. None of the recommendations below should be read as a recommendation for training a machine-learning classifier or calculating expected monetary value. If you need probabilities, model fitting, or numerical optimization, choose analytical software for that task. For the terminology, see what a decision tree is.
Free Decision Tree Makers: Limits and Best Uses
| Tool | Free-plan boundary | Editing and sharing | Useful starting situation |
|---|---|---|---|
| draw.io | Free web and desktop editor; some host integrations are paid | Editable files; real-time editing in supported Google Drive or OneDrive workflows | You need a reusable file and control over storage |
| Lucidchart | 3 editable documents, up to 60 shapes per document | Browser diagramming and collaboration; check plan-specific export permissions | You have one small formal tree |
| Miro | 3 editable boards | Collaborative whiteboard; board access and advanced features depend on plan | The team will debate the branches together |
| Canva | Free templates and assets; premium content and features require payment | Visual design and presentation workflow | The logic is settled and the tree belongs on a slide |
| Figma / FigJam | Starter file and team permissions apply; do not assume unlimited shared FigJam files | FigJam shapes, connectors, and live collaboration | Your reviewers already work in Figma |
| CodePic | Browser whiteboard and editable decision-tree example | Manual editing, PNG export, read-only link sharing | You want a sketch-like tree without a team workshop setup |
| MindMeister | 3 maps; file exports require a paid plan | Collaborative branching maps | A hierarchical outline is enough |
Limits were checked against official sources on October 24, 2026. These are different kinds of free access, not equivalent unlimited plans. A map cap, an object cap, and a board cap constrain different parts of your work. The comparisons describe documented capabilities rather than results from a timed seven-product test.
draw.io: A Free Editor with Editable Files
Choose draw.io when the tree needs to survive outside a particular cloud workspace. You can keep the editable source and produce an image for a document. Its standard shapes and connector labels are sufficient for a question-and-answer tree; a dedicated “decision tree” mode is not required.
A useful layout has the root question at the top, Yes consistently on one side, and outcomes aligned along the bottom. The editor includes automatic layout options, but recheck every branch label after rearranging a tree. A layout engine can improve spacing without understanding whether the rule itself is correct.
Collaboration deserves a correction to a common assumption: draw.io does support concurrent editing through supported Google Drive and OneDrive workflows. That does not mean every local file or Git-backed workflow automatically becomes a live meeting. Choose the storage integration first, then confirm that both editors can join the same file. The draw.io vs. Lucidchart comparison covers this storage-versus-workspace choice in more detail.
Lucidchart: Small Formal Trees in a Shared Workspace
Lucidchart is a candidate when a small tree needs to sit beside other business diagrams and be reviewed in a browser. The free allowance of three editable documents and 60 shapes per document makes diagram size a concrete selection factor.
Count the questions and outcomes, not just the questions. A full binary tree with 20 decision nodes has 21 leaf outcomes: that is already 41 nodes before optional explanatory shapes. Object-counting rules vary by product, so this arithmetic is a planning example rather than an assertion about how every connector is billed.
Before expanding a policy tree, try the export you actually need and invite a reviewer with the intended permission. A diagram that fits the free canvas can still fail your delivery requirement if the required file format or access level belongs to another plan. Do not assume advanced diagramming features are included just because the basic shapes are available.
Miro: Build the Rules during a Workshop
Miro fits a meeting where the team is still deciding what each branch should mean. Keep the evolving tree, unresolved questions, and supporting notes on the same board. Separate “the current rule” from proposed alternatives so reviewers do not mistake a discussion sticky note for an approved outcome.
The free plan includes three editable boards. That is a workspace limit, not a promise that all facilitation, export, or access-control features are free. If a board already contains a large workshop, put the tree in a clearly labeled area and keep its reading direction consistent.
You do not need a separate tool just to replace rectangles with diamonds. You may need a separate deliverable if the final policy must be read as a compact page by people who will never join the board. Prepare that page before treating the workshop as finished.
Canva: Put a Stable Tree into a Presentation
Canva is useful when your destination is a slide, handout, or report and the branch structure is already agreed. Its decision-tree templates provide a starting visual composition, with free and premium assets distinguished in the design workflow.
Choose a template with enough room for your longest question and the number of outcomes you actually have. A symmetrical example with four leaves is a poor starting point for a policy with seven uneven paths. Forcing the policy into a decorative arrangement can hide an exception or make the text too small.
Test one structural change before committing: add a fifth outcome and move a question to a different parent. This tells you whether the chosen design remains manageable. The relevant issue is the cost of future revisions, rather than whether the first screenshot looks finished.
FigJam: Keep a Design Review in the Existing Workspace
Use FigJam when the tree explains a user journey, permission choice, or product interaction that designers will review beside existing work. Shapes, connectors, comments, and collaborative editing support that conversation.
Figma Design and FigJam are different products. A component layout in a design file is not the same thing as a branching diagram on a FigJam board. Also check Starter file limits and team permissions instead of assuming all shared FigJam files are unlimited.
For a product rule, label outcomes with behavior: “Show account recovery” is more actionable than “Failure.” Reviewers should be able to follow a branch from a concrete user state to an observable screen or action. If the discussion concerns the sequence after an outcome, a flowchart may express it better than a tree.
CodePic: An Editable Tree on a Hand-Drawn Whiteboard
CodePic offers an editable decision-tree example on an infinite whiteboard. Start from its existing structure, change the questions and outcomes, and adjust positions and connectors manually. The template supplies a layout; it does not continuously rebalance every future branch.
You can export a PNG or share a read-only link. Read-only sharing is useful for review, but it is different from several people editing the same tree live. An optional MCP workflow can provide a starting diagram from an AI assistant; the assistant's availability and charges are separate from editing the whiteboard.
Use the example below for the support-routing exercise below. It is a starting canvas, not an executable support policy or a trained decision model.
MindMeister: A Branching Outline before Formal Notation
MindMeister is useful if your first problem is organizing questions into a hierarchy. Its Basic plan includes three maps and collaborators, while file exports require a paid plan. That makes it a different fit from a local-file editor when handing off an image or document is mandatory.
A mind-map branch is also not automatically a Yes/No connector. Write the answer into an intermediate branch label or make the child text explicit enough that the relationship cannot be misread. Do not rely on color to distinguish decisions from outcomes.
If the tree needs repeated checks, reconverging branches, or loops back to an earlier step, switch to a flowchart rather than forcing the process into a hierarchy. Mind maps organize related ideas; a decision tree needs every path to express a rule.
Worked Example: Route a Support Request
Here is an illustrative tree with three questions, four outcomes, and six labeled branches. It is a reusable drawing exercise, not a production incident-response policy.
| Question | Yes branch | No branch |
|---|---|---|
| Is this an account-access issue? | Can the user sign in? | Are multiple people affected? |
| Can the user sign in? | Contact the workspace permission owner | Start account recovery |
| Are multiple people affected? | Open a team incident ticket | Open an individual support ticket |
Draw the first question at the top, the next two questions below it, and the four outcomes on the final row. Label every connector Yes or No. The word “Yes” belongs on the edge leaving the question, not inside a distant outcome box.
Trace two example inputs. A person who cannot sign in should reach account recovery. A non-access problem affecting several people should reach the team incident ticket. If either route requires guessing, revise the labels before changing colors.
Now introduce a realistic revision: some users do not know whether anyone else is affected. Add an Unknown branch from the third question to “Ask for impact details.” This is a third answer, not a second No branch. It also exposes a scope choice: is collecting those details part of this tree, or handled by a linked procedure?
Common Mistakes When Choosing a Free Tool
Counting only the first version. A four-outcome template can be easy to fill and difficult to revise. Add an exception and move one branch before choosing.
Treating sharing as editing. A public view link, an invited editor, and an exported PNG solve different needs. Check the recipient's exact task.
Drawing questions that overlap. “Urgent?” and “Important?” have no shared definition by themselves. Define the threshold or use observable conditions so two readers reach the same result.
Expecting the diagram to execute the policy. A drawing does not enforce permissions, send tickets, or validate a classifier. Link the final diagram to the procedure or software that performs those actions.
Choosing before checking the final output. Export a small sample at its intended display size. Branch text should remain readable in the actual wiki column or slide, and the editable source should have a clear owner for future corrections.
Related Reading
- Decision tree definition and examples — distinguish questions, branches, and outcomes.
- How to make a flowchart — use a process diagram when paths loop or reconverge.
- Free flowchart makers — compare tools for more general workflows.
