Persona
Identity, goals, behavioural state and accumulated memory.
115 Studios / Autonomous mobile-user platform
HAB1T
Persistent synthetic users. Real mobile behaviour over time.
HAB1T is an on-premises platform for running persistent synthetic mobile users across Android environments. Each persona carries its own goals, behavioural state and memory, acts through real applications, observes what happened, recovers from blockers and continues from the state it has built up.
Deployment readyAndroid execution8 selectable personasLocal ControllerAuditable evidence
Self-contained Windows deployment / Local-first operation / Physical Android enrolment
PERSONA / DEVICE / STATE115 STUDIOS / HAB1T
What HAB1T is
A user model that keeps living after the first action.
Most mobile automation begins with a fixed sequence: tap here, type this, wait for that, assert the result. HAB1T starts somewhere else.
A HAB1T persona has an identity and a changing internal state. It can be given goals, make a decision about what to do next, pass that intent to an Android executor, inspect the result and use the evidence to shape what happens afterwards. Work can be interrupted and resumed. Blockers can change the plan. Previous activity can matter later.
That makes HAB1T useful where the question is not only “did this scripted flow pass?”, but “how does this software behave when a persistent user keeps interacting with it over time?”
The HAB1T loopBehavioural intent stays separate from device execution
Decision
Choose intent, priority and the next bounded action from current evidence.
Android
Translate intent into device interaction through a replaceable executor.
Observe
Inspect the outcome instead of assuming the action succeeded.
Continue
Update state and memory, recover or replan, then carry the session forward.
Inside HAB1T
Built around behaviour, not a macro recorder.
The current platform combines persistent personas, autonomous session logic, Android execution, a local Controller and machine-readable evidence into one system.
Persistent personas
Eight selectable personas have been built and validated with individual goals, behavioural state and cross-session memory.
Autonomous sessions
Sessions can choose between goals, interrupt and resume work, replan from evidence and recover when the expected route stops working.
Android execution
Behavioural intent is kept separate from the Android executor so the same persona model is not welded to one device or interaction path.
Physical device identity
Physical Android enrolment preserves device identity, detects mismatches and fails closed rather than silently treating changed hardware as the same slot.
Local Controller
The self-contained Windows deployment provisions the HAB1T Controller, local TLS, persistence, ADB tooling and the services needed to coordinate registered devices.
Evidence first
Decision, observation and outcome evidence is captured so a session can be inspected afterwards instead of disappearing into an opaque automation log.
Production capability
A complete system, from persona to deployed device.
HAB1T combines persistent persona runtimes, Android execution, Controller services, physical-device enrolment, deployment tooling, automated recovery and auditable evidence into one operational platform.
The Windows deployment path has completed clean-machine and real-device acceptance, giving the platform a defined route from installer to Controller provisioning, device enrolment and sustained autonomous execution.
PERSONAS8 distinct profilesSelectable identities with persistent behavioural state.
BEHAVIOURReplan + recoverInterrupt, resume and adapt from observed results.
ANDROIDEmulated + physical executionAndroid actions have been exercised beyond emulator-only work.
DEPLOYMENTSelf-contained Windows pathInstaller, Controller provisioning and device-enrolment tooling.
AUDITMachine-readable evidenceDecisions, observations, outcomes and deployment proof.
ACCEPTANCEClean Windows + real Android verifiedEnd-to-end deployment and physical-device path accepted on target hardware.
Where it applies
Software that needs more than one perfect user.
HAB1T is most useful when behaviour over time matters: different users, changing state, interruptions, imperfect routes and enough evidence to explain what happened.
Mobile QA & regression
Exercise applications through persistent user profiles instead of relying only on short deterministic paths. Useful for finding state-dependent behaviour that appears after earlier actions have changed the user or the app.
Long-duration device testing
Run user-like activity across longer sessions where pacing, interruption, recovery and accumulated state matter more than completing one test case as quickly as possible.
Device-farm autonomy
Use a Controller-and-slot architecture to coordinate synthetic users across registered Android execution targets while keeping persona behaviour separate from the device layer.
Cybersecurity validation
Observe how modified or instrumented mobile software behaves under varied user behaviour, with a traceable record of the decisions and outcomes that produced a result.
Accessibility & usability testing
Run repeatable behavioural profiles through longer interaction histories while retaining the evidence behind each automated session for inspection and comparison.
Behavioural scenario testing
Place distinct persistent personas into the same starting conditions and observe how memory, priorities and previous outcomes produce different routes through the same software environment.
The research behind it
HAB1T was not built in one pass.
HAB1T grew through a long series of controlled experiments around behavioural identity, memory, decision continuity, interruption and recovery, identical-world persona divergence, device compatibility, physical-device lifecycle, guardrails, deployment and evidence capture. That research is now embedded in a finished platform rather than remaining a collection of isolated experiments.
One of the important findings was simple but useful: when distinct personas are placed into the same world, they do not have to collapse into the same sequence of actions. Their goals, discretionary behaviour and resulting state can diverge while still operating through the same device environment.
H1
A different testing surface
Scripted automation still has a job. HAB1T solves a different one.
Exact deterministic tests are excellent when the expected sequence is already known. HAB1T is aimed at the messier space around them: persistent state, competing goals, unexpected blockers, behavioural variety and the consequences of what happened earlier.
The goal is not to make testing less precise. It is to make a different class of behaviour testable and inspectable.
Product state
Finished platform. Built for controlled, on-premises deployment.
HAB1T is a complete 115 Studios platform with tested software, deployment tooling, physical Android enrolment, persistent personas and a documented evidence trail. The product is designed to run locally under operator control rather than depending on a hosted account or external cloud service.
This page explains the product, the architecture behind it and the classes of mobile testing where persistent synthetic users provide a different capability from conventional scripted automation.
STATECOMMERCIAL RELEASE / DEPLOYMENT READY
PRIMARY EXECUTIONANDROID
CONTROLLERLOCAL / ON-PREMISES
PERSONAS8 PRODUCTION PROFILES
WINDOWS DEPLOYMENTPACKAGED + HARDWARE-ACCEPTED
PHYSICAL ANDROIDREAL-DEVICE ENROLMENT VERIFIED