The branch stays closed while the hostile-faction condition remains true.
Conversations
Hear the conversation before the audience does.
Explore branching dialogue in the context of the characters and world it belongs to, while every result remains a draft experiment.
Planned for public betaApprentice+
Dialogue testing is intended to help Game Masters prepare difficult NPC exchanges, writers explore alternative responses and narrative designers inspect branches before anyone encounters them in a finished work.
Mimir's authored Dialogue graphs support speakers, choices and condition-gated branches. The public Proving Grounds workflow will bring that structure into a clearer testing experience.
- Try a different response
- Inspect the branch it opens
- Keep experiments separate from canon
Conceptual branching conversation
Scene contextCommander Arlen guards Greywatch's sealed gate.Known state: House Vael is hostile · northern road fallen
Commander Arlen
“No one enters while Vael banners stand beyond the ridge.”
The conversation opens a cautious route because the envoy knows about the old passage.
The hostile path ends the exchange and creates a possible consequence to test.
Illustrative dialogue. It does not represent a live Proving Grounds session or built-in Mimir world.
Quests & choices
Find the broken branch before your players do.
Test objectives, prerequisites, choices, consequences, failure paths and alternative outcomes while changes are still cheap.
Conceptual quest-condition tester
Quest
Reach Greywatch
Change the conditions to see which branch remains possible.
StartReach the outer wall
→Available branchRequest entry from Commander ArlenThe gate is open and its commander can respond.
→OutcomeEnter under guard
The main-gate route is currently available.
World state
Choices matter when the world remembers them.
Variables and conditions describe the facts that can change during a test, so later branches can respond to what happened earlier.
In developmentWorldsmith+
World state is different from permanent canon. During a test it can hold a temporary value such as an open gate, an alliance, a living commander or a fallen road, then let the next part of the experiment react.
Mimir's authored variable system supports typed global and scoped values. The full public workflow remains under development, and this demonstration deliberately uses simple on/off states.
Conceptual state board
Change the world. Watch the consequence.
Current consequenceThe envoy enters under guard, but Greywatch remains isolated.
Arlen is alive and the gate is open. The road is still fallen and House Vael remains hostile.
This is an explanatory model of changing state, not the actual Mimir interface.
Sessions, systems & simulation
Test more than a single branch.
Use the same connected world to prototype a conversation, inspect a quest, explore a session or challenge a lightweight system idea.
Choose a testing lens
One world. Different experiments.
Character response and branching dialogue
Try alternative lines, choices and condition-gated responses against established character and world context.
UsesDialogue graphSpeakersChoices
Objectives, branches and consequences
Walk a quest graph, inspect changing variables and find unreachable or incomplete routes.
UsesQuest graphConditionsState
Rules-aware play and session rehearsal
Select governing Rule entries, take a role and explore a persistent local session with AI-controlled seats where appropriate.
UsesAuthored rulesRolesVisible app-rolled dice
Lightweight scenario and rules experimentation
Probe how defined rules, choices and state interact without claiming to replace a game engine or real playtesting.
UsesRule entriesScenario stateObserved outcomes
The selector explains intended testing modes. It is not a promise that every mode has the same implementation or release status.
The testing environment
Where the creator runs an experiment using the world they built.
The facts that can change
Variables, relationships and conditions that shape the current test.
The system processing change
Stepping through authored logic, state and outcomes so behaviour can be inspected.
In development
A real foundation, still being brought to public-beta quality.
Current product work includes local persistent sessions, selected governing Rule entries, human and AI-controlled seats, transcripts, visible app-rolled dice, Quest and Dialogue simulation, state inspection and replay export. The complete customer-facing Proving Grounds experience remains planned for public beta.
The creator decides
An experiment can inform the world without taking control of it.
Run the test, inspect what happened and learn from it. The creator decides whether to revise the world, discard the result or deliberately carry an idea forward.
Conceptual creator judgement
Experimental result
The forgotten passage makes the siege too easy to bypass.
This result reveals something useful about the scenario. It has not changed the world's source of truth.
No decision has been made. The result remains an experiment.
These are explanatory choices, not a claim that Mimir currently has a one-click “accept into canon” workflow.
Planned for public beta
Local experiments first. Hosted participation when the test needs people.
Apprentice+ is planned to create and run Proving Grounds locally.
Worldsmith+ is planned to host live collaborative Proving Grounds. Invited Core users can participate within the host's allowance.
Put the world into motion
Build it. Test it. Decide what survives.
Explore every Mimir capability or join the beta and help shape the complete Proving Grounds workflow.