E:Voice is a voice-first messaging app built for AI-native work. Tool calls, active thinking, and backend execution become organized thought boxes instead of chat spam — with full detail in text and clean summaries in voice.
AERIS OS powers E:Voice native AI agents that can discuss bugs, infer intent, route work, judge whether responses are complete, and return useful next steps. When you're busy, Frank can review an agent's summary, decide whether the work is moving or stuck, and nudge the right agent back into action. Future Build Factory integration means AERIS agents can build websites, apps, and high-turnaround modular AI apps in a safe environment.
Internal and external agents can share and write to an Obsidian-style memory vault. App information stays stored locally and securely on your device; it is not saved to our servers.
Save provider keys in one encrypted place. The API Key Vault works like a mobile wallet, releasing keys to agents only when they need access and protecting vault entry with your Android screen-lock security.
Assign seats so one Foreman can direct work, or send a handoff like: "Spin, nice audit on that bug. Send the work order to Charles for implementation." You can manually pass 5, 10, 15, or 20 messages to another agent with a short discuss-before-execution header.
E:Voice keeps the conversation human while letting the AI work stay visible. Tool calls, active thinking, terminal commands, status updates, and handoffs become organized boxes inside the app instead of flooding the transcript.
Typed chat can carry the complete reasoning-safe response, links, files, settings, and next steps without cutting the agent short.
Voice gives you the useful result first. When you pick up a thread, E:Voice can read the last message out loud and continue from there.
Working notes and tool status are visible, but separated from the conversation in expandable boxes so the chat remains readable.
Quick-action hard stops send immediate bridge commands to halt execution and can also send /stop in chat. The E:Voice bridge requires voice-command prompts to get a discussion response before execution, so your TV cannot accidentally tell your agent to nuke the project.
Frank is the onboard ChatGPT expert for people who know the feeling, goal, bug, or product direction but not yet the exact prompt. He questions, clarifies, shapes the brief, and routes the intent to the right internal or external agent.
Voice or text captures the messy version while the idea is fresh.
Clarifying questions turn desire, frustration, or design taste into a usable brief.
Select an agent and forward the last 25 messages with a generated discuss-first prompt.
Hermes, Codex, OpenClaw, AERIS OS, or a customer runtime gets only the context it needs.
The result, summary, files, and next decision come back to the E:Voice thread.
Generic human messaging apps with AI bolted on collapse AI work into a single noisy thread or force you to switch between threads. E:Voice lets you save tasks and projects, then stores context in that project memory. You keep one thread, while any agent can refer to the context inside the project.
Related chats, decisions, notes, screenshots, files, and handoffs stay grouped by workstream.
Hermes, OpenClaw, Agent Zero, AERIS agents, customer servers, and future agents can be routed through explicit bridge surfaces.
Oversight can notice stalled agents, ask for missing completion, and keep long-running work from disappearing into the chat archive. "Wrap this task in a Pulse" means the work does not stop until the objective is achieved.
Good ideas and bugs can move immediately into implementation and resolution. E:Voice is used every day to optimize how it streamlines your workflow and looks good doing it, hands free. There is no limit to how far E:Voice can grow.
E:Voice bridge packages make setup easier for customer-owned Linux runtimes. Start with the Hermes integration package, or connect OpenClaw and Agent Zero with the same setup-code pairing model. The packages are Linux-optimized and verify plugin downloads before install.
Installs the E:Voice Hermes bridge helper, verifies the current runtime plugin package, and prepares Hermes for E:Voice pairing and delivery routing.
curl -fsSL https://www.emergencerising.com/evoice/integrations/hermes/install.sh | bash
Downloads, verifies, backs up, and installs the current E:Voice OpenClaw bridge extension into the user-owned OpenClaw extension directory.
curl -fsSL https://www.emergencerising.com/evoice/integrations/openclaw/install.sh | bash
Downloads and verifies the Agent Zero E:Voice plugin package. Because Agent Zero installs vary, pass --plugin-dir or set AGENT_ZERO_PLUGIN_DIR to install automatically.
curl -fsSL https://www.emergencerising.com/evoice/integrations/agent-zero/install.sh | bash -s -- --download-only
These packages are maintained for E:Voice integration. They do not imply upstream official support from Hermes, OpenClaw, or Agent Zero unless those projects separately list E:Voice as supported. The goal is a clean customer install path without Dan-specific infrastructure.
Connect it to your preferred agents and voice providers, then start working with your stack while you're away from your desk. Want to put the phone down when you get home? Install AERIS OS and keep the same ecosystem synced on your PC.
E:Voice does not come with external compute. Log in to your ChatGPT account for easy OAuth, or add your API key providers for local AERIS AI agents. The AERIS AI inside E:Voice is intentionally lightweight. Power Frank and the rest of your AI team with your ChatGPT Plus account, then keep that team dedicated to moving your work forward.
Pre-launch download access remains account-gated where enabled. Paid restrictions are not live yet, and E:Voice billing will be handled through Google Play when the app fully launches. Cancel anytime.