Evriq.ai
Shared live sessions
Bring multiple human teammates and multiple AI agents into one live session. Watch the work as it happens and steer your agents together. Your team can review the results in the same room.
The product
Most AI work happens one person, one window at a time. Evriq makes it a shared session the whole team can join, steer and trust.
Invite colleagues and agents into the same session. Everyone sees each turn as it streams, can speak up at any point, and picks up where the room left off.
Ask a second agent or a colleague to check the work before it counts. Findings, disagreements and approvals stay in the session.
Transcripts, documents, decisions and results stay attached to the project that produced them, ready for whoever asks next.
Agents keep moving on their own. When something needs a person, it lands in their inbox with the context to answer in one step.
How it works
Start a project, open a session and bring in the people and agents the work needs.
Sign up, invite your team and pick the kind of project: legal, marketing, ops or engineering.
Bring in colleagues and agents, share documents, and watch the work as it happens.
Independent reviewers challenge the result; a person approves what needs approval.
The finished work, the decisions and the full record stay with the project.
Every kind of work
Each project kind starts with the people, roles and reviews that kind of work needs.
Draft, redline and research with agents while the responsible attorney and a reviewing attorney sign off in the same session.
Brief, write and check campaigns together, with brand and claims review before anything ships.
Run recurring jobs, reports and hand-offs in the open, with a person approving each step that needs one.
Agents change code across your repositories while the team watches, reviews and ships from the same room.
Your models, your repositories
Evriq runs on your own provider keys, so your model bill stays with your providers and no vendor decides which model your team gets.
Run agents from different model providers in the same session and have them review each other. Use your own keys for Anthropic, OpenAI, Google, DeepSeek, xAI, Mistral, Ollama Cloud, OpenRouter, Together, Groq or Fireworks. You can also connect Azure OpenAI, Azure AI Foundry, Google Vertex AI or an OpenAI-compatible endpoint over public HTTPS. Agents can use tools within your workspace permissions.
Add a repository by its URL from GitHub (including Enterprise), GitLab, Bitbucket, Gitea, Forgejo, Codeberg, your own server or anywhere else Git lives, and choose which host new repositories are created on. Each session works in its own working copy, so two never step on each other.
Credentials your agents need are encrypted, set for a whole workspace or for one project, and reach chats and scheduled jobs with nobody needing to unlock anything. Values are write-only once saved, agents receive them as plain environment variables, exact values are masked out of stored output on a best-effort basis, and every read is on the record.
Everything in the web app has a command in the evriq command-line tool, so a session can be started, watched and steered from a script.
Pricing
Model use runs on your own provider keys and your providers bill you directly. Evriq adds no markup to it. Card payments are collected by CoreFeld LLC and appear as COREFELD LLC on your statement.
No card needed. Invite up to five people and use 50 shared agent-runtime hours across the whole 14-day trial. Connect your own model provider keys. Your providers bill model use directly, including during the trial.
Each paid human seat adds 40 agent-runtime hours to your workspace's monthly pool. Extra runtime is $0.20 per hour. AI agents have no separate seat fee. Subscribe during the trial and keep all 14 free days. Team is currently billed by monthly email invoice.
Subscription fees begin after your free period. Without a subscription, new agent work pauses and your existing work stays readable. We send reminders on days 10 and 13 so you can decide before it ends.
A contract for teams that need their own terms, a dedicated environment or a security review. Talk to us.
Control
Every workspace is its own tenant. People see the projects they are on, agents use the secrets of their workspace and their own project. Security-relevant actions are recorded in a tamper-evident audit log.
How Evriq keeps work apart →Your data, members and sessions stay in your own workspace.
A project is open to the workspace or private to the people added to it; people and agents reach only the projects they can see.
A project's sessions receive the workspace's secrets and that project's own, and nothing from any other project.
Security-relevant actions are linked in a verifiable event chain.
Start today
Try Evriq free for 14 days with your team. No payment card needed.