Not gateway-first
E:Voice is the control surface, not just a remote dashboard. It lives where you are.
E:Voice is not a chatbot skin and it is not trying to replace every advanced agent stack. It is a mobile-first cockpit for using agents, projects, tools, bridges, and proof without living in a terminal.
E:Voice is the control surface, not just a remote dashboard. It lives where you are.
Speak tasks, review status, and keep work moving without sitting at a terminal.
Frank scopes the request. Foreman plans the work. Specialists handle focused execution.
Daily logs, sessions, project notes, core memories, tags, links, search, and cleanup controls stay visible to users and agents.
Meaningful actions should leave a trace: who acted, what ran, what changed, and what proof came back.
Android stays light. Builds, logs, artifacts, and proofs run where they belong.
Native voice-to-text and TTS options keep the phone responsive without forcing every voice action through a cloud round trip.
Use E:Voice alongside OpenClaw, Hermes, custom HTTP runtimes, or other approved agent systems.



E:Voice is the phone and voice cockpit. AERIS OS is the desktop and browser mission-control layer for agents, machines, dashboards, files, shared memory, calendar, email, docs, and stack orchestration.
Talk to agents, route work, review proof, and keep projects moving away from the desk.
Connect the dashboards, machines, files, and agent systems that power the stack.
Carry project facts across Hermes, Codex, Claude, AERIS agents, and E:Voice sessions.
E:Voice at $5/month and AERIS OS at $15/month are designed to work together without replacing your stack.
A careful comparison, not a takedown. Each category is useful. E:Voice focuses on making agent work mobile, visible, and project-centered.
| Category | ChatGPT-style chat | Terminal or gateway agents | E:Voice |
|---|---|---|---|
| Primary interface | Conversation window | Terminal, server, dashboard, or IDE | Android-native mobile cockpit |
| Voice-first workflow | Often available, usually conversation-focused | Usually not primary | Designed around speaking tasks and reviewing work from the phone |
| Mobile-native experience | Strong for chat | Often PC or server-first | Built for mobile approvals, status, projects, and routing |
| Project cockpit | Usually not a full project layer | Powerful, but often fragmented across tools | Projects, agents, bridges, receipts, updates, and build proof in one lane |
| Agent orchestration | Limited or hidden | Deep, but technical | Frank intake, Foreman planning, specialists, and bridge routing |
| Memory visibility | Varies by platform | Depends on the stack | Obsidian-style memory vault with layers, search, tags, links, metadata, and cleanup controls |
| Tool receipts and proof | Often conversational summaries | Logs and artifacts, if wired | Action receipts and proof surfaces keep work inspectable |
| Local desktop power | Not the primary model | Strong when configured | Desktop Companion is an early-stage local power layer for approved tools and proof |
| Works with existing stacks | Usually through plugins or APIs | Native to the stack | Designed to complement OpenClaw, Hermes, custom HTTP bridges, and Codex lanes |
| Beginner friendliness | High | Lower | Start with Frank and Foreman, then grow into advanced routing |
| Power-user depth | Good for reasoning and drafting | Very high | Mobile cockpit for advanced execution, not a replacement for every stack |
E:Voice keeps the control surface simple while leaving room for deeper execution lanes underneath.
E:Voice is already a working Android cockpit and keeps growing. These are the surfaces the product story now leads with.
Start with a voice note, scope the project, route the build, and review proof when execution returns.
Check status, unblock tasks, and approve the next lane from Android.
Frank captures intent and Foreman turns it into structured work.
Use E:Voice as the intake and approval surface for source work.
Keep builds, logs, updates, receipts, and release proof attached to the work.
Bridge to OpenClaw, Hermes, or custom runtimes without making them the only front door.
E:Voice should show what agents can do, what they did, what they remembered, and which bridge handled the work.
Know which agent owns the conversation, the plan, and the execution lane.
Approved tools and bridges should be explicit, reviewable, and scoped to the project.
Every meaningful action should produce a trace users can inspect.
Users and agents can search linked daily logs, sessions, project notes, and core memories instead of relying on short-lived context.
Project bridge controls keep OpenClaw, Hermes, Codex, Companion, and custom runtimes understandable.
Artifacts, logs, update delivery, and proof should stay attached to the work.
E:Voice Companion is in the early stages of development. It pairs your phone with your computer, exposes approved tools, manages local projects, and brings desktop proof back into the mobile cockpit without asking users to install a full agent gateway.
Connect a phone to an approved desktop workspace.
Keep project context close to the machine that can execute it.
Expose only the tools and files that should be available.
Route coding work through the desktop when the lane is approved.
Reviewed skill packs can extend local execution safely.
Local actions should return proof to the mobile cockpit.
Use desktop power without giving up review and approval.
Companion should work alongside advanced stacks instead of hiding them.
The current photo shoot shows the product as it is used: chat, memory, docs, skills, voice settings, agent creation, image generation, appearance, themes, and stack-aware agent cards.

Landscape chat shows the core idea: color-coded agents, live connection, Frank intake, bottom navigation, and a focused project conversation.

Agent cards expose roles, custom identities, recent activity, and visual colors for different stacks.

Obsidian-style memory organizes daily logs, sessions, projects, core memories, links, tags, and searchable context.

Local docs bring notes, to-do lists, prompts, reports, and agent-generated documents into the app.

Provider keys, services, OAuth image generation, navigation preferences, and local app settings stay visible.

Vosk offline STT and local TTS choices support low-latency voice loops on Android.

Saved voices, local encryption, preview controls, and provider selection make voice output configurable.

Customize agent card colors for Frank, AERIS, OpenClaw, Hermes, and choose wallpapers for the app.

Create native AERIS agents or add external runtimes with role templates and identity controls.

Agent setup can upload or generate profile pictures using OAuth image generation and editable prompts.

Skill cards keep available agent capabilities organized by need, status, and workflow.

Specialists can carry focused conversations while staying inside the same mobile cockpit.

Custom agents keep their own identity, wallpaper, color, and execution context.

Specialists like Torx can hold UI direction, critique, and design implementation context.

Chat can carry image-heavy planning, references, and proof instead of staying text-only.

Frank can keep detailed product, build, and planning conversations readable on mobile.

System identity, versioning, preview status, and execution boundaries are visible inside the app.

Settings expose defaults, voices, models, and app behavior without hiding control in backend config.

Bottom tabs, builder profile, runtime, and appearance controls make the app adjustable to the workflow.
Talk to Frank. Let Foreman organize the work. Choose where execution happens: Android, Build Factory, Codex, Companion, OpenClaw, Hermes, AERIS OS, or a custom bridge. Access codes can unlock trial, creator, tester, partner, or earned use windows.