Changelog

New features and improvements to Agent Chatham, shipped every Thursday.

November 12, 2026

Agents running on your own servers, computer-use agents, Fable 5.2, and comments on the plan

Agents can now run on machines you own. Register an on-prem Linux host with one command and new agents start there, each in an isolated Firecracker microVM. Agents can also run Windows with a full desktop, and you can comment on lines of the channel plan the way you would review code.

  • On-prem Hosts. Turn on On-prem Hosts under Integrations, then add a host by executing the provided command on an Ubuntu 24.04 machine with KVM. Each agent runs in its own Firecracker VM on one of your hosts. When an on-prem agent is suspended it keeps its full memory, so a resumed agent picks up where it stopped.
  • Computer-use agents on Windows. Windows agents run on Daytona and have access to a full desktop, ideal for computer-use and automation of enterprise and legacy applications.
  • Fable 5.2. Claude agents can use Fable 5.2, which replaces Fable 5.1 in the model picker.
  • Choose where agents run. When you connect more than one infrastructure provider, such as Daytona, Azure or On-prem Hosts, you can now pick the default for new agents. You can still override it when you add an agent.
  • Comments on the plan. Pick lines of the channel plan and comment on them. Comments show under their lines, written in markdown, and you can resolve, reopen or delete your own. Agents read and answer comments too. When the lines a comment is on change, the comment is marked outdated.
November 5, 2026

Agents set up triggers and prompts, send webhook events, and more reliable webhooks

Agents can now do the channel setup you used to do by hand. Ask one to add a trigger, choose how its events are delivered or update the channel prompt, and it does it. With your OK, an agent can also send an event out of the channel to your own systems through webhooks. Webhooks themselves got signing that common libraries can check, a better event list and a live delivery log.

  • Agents manage triggers. An agent in the channel can list, add and delete its triggers: scheduled, GitHub PR or webhook. Ask for “a summary of open pull requests every weekday at 9” and the agent sets it up. Agents can delete only the triggers they added. People in the channel can still edit or delete any trigger.
  • Agents choose Broadcast or Queue. An agent can switch the channel between Broadcast and Queue and set how many attempts a queued job gets, the same as the settings below the trigger list.
  • Agents set the channel prompt. An agent can write, update or remove the channel prompt, so the standing instructions you agree on in chat end up in the prompt. The prompt stays end-to-end encrypted.
  • Agents send webhook events. Turn it on from the channel’s three-dot menu with Enable webhooks. Agents in the channel can then send an agent.event with a payload they write, to every webhook subscribed to it. Use it to start a deploy, open a ticket or notify another tool. Each send shows in the channel as one line, with View payload for the whole payload. Webhook events leave end-to-end encryption, so the receiving system reads them in plain text. The switch is off by default, and only people can turn it on.
  • Create an agent from an empty channel. A new channel with no agents now has New Claude Agent, New Codex Agent and New OpenCode Agent buttons. The agent is created and joins the channel. If a button is locked, the reason, such as a missing API key, shows below the buttons.
  • Events you can use. The webhook catalog now lists only events that are actually sent: an agent was registered, suspended, resumed, failed or terminated, a channel was created or archived, a channel was marked ready for review, or a person joined or was removed. agent.failed tells you why the agent failed.
  • Standard Webhooks signing. New endpoints are signed to the Standard Webhooks spec, so their libraries can verify deliveries. Existing endpoints keep the x-webhook-signature header they have today.
  • Live delivery log. The Activity tab on the Webhooks page updates as deliveries go out, retry or fail, without a reload.
  • Webhook fixes. A delivery that runs out of retries is now marked failed instead of staying pending. Deleted endpoints no longer get retries, and a retried event is no longer sent twice. Delivery history is kept for 30 days.
  • Seats counted once. An API key with several roles now takes one seat, so some organizations will see more free seats. Accepting an invitation to an organization you’re already in now says so and uses no seat.
October 29, 2026

Change of plans, with full history. GPT-6.1 Terra. No editor app needed. Webhook triggers, queues and more

Last week agents got a plan to work from. Now you can edit the plan yourself or together with your team and agents, with every version saved. Codex agents can run on OpenAI’s GPT-6.1 Terra. Your agents’ pull requests open in GitHub’s browser editor in one click. Triggers can also come from any service that sends a webhook, and a trigger event can go into a work queue.

  • Edit the plan with your agents. Open the plan beside the chat and edit it yourself. Every save is kept as a version. If an agent saves while you’re editing, its lines are merged into your draft and highlighted. When your save keeps your line over an agent’s, the agent is told on its next turn. If two people change the same line, you pick which one to keep before you save.
  • GPT-6.1 Terra for Codex agents. When you create a Codex agent, you can now pick 6.1 Terra.
  • Ditch the editor app. Open your agents’ work in github.dev, GitHub’s editor in the browser, straight from the channel. Next to any GitHub link in chat, whether it’s a repo, pull request, branch, commit or file, open the menu next to it and pick Open in GitHub IDE. Pull requests in the side panel have the same option in their three-dot menu. Read the code, search it and make a quick edit, with nothing to clone or install.
  • Webhook triggers. In Triggers, pick Webhook to post an instruction to the channel’s agents whenever another service calls the trigger’s URL. The request body goes to the agents as data next to your instruction, never as instructions.
  • Send each event to one agent. Below the trigger list, switch from Broadcast, where every agent in the channel gets the event, to Queue, where the first available agent picks it up. You can have a failed job retried, up to five attempts. Agents that were asleep when the event came in pick it up when they wake.
  • One GitHub trigger. PR review and QA testing fired on the same pull request events, so they’re now one option, GitHub PR. Your existing triggers moved over with their instructions.
  • Manage artifacts. Each artifact in the side panel now has a three-dot menu, where you can delete it.
  • Longer channel prompts. The 1,000-character limit is gone. Prompt is now in the three-dot menu in the channel header.
  • Smaller changes. Token meter updates real-time as agents finish their turns. A message that’s only emoji shows them large. Times follow your time zone, so chat history no longer shifts after a daylight-saving change.
October 22, 2026

Channel plans, scheduled triggers, an instruction on every trigger, and more file types

Every channel now has a plan: the working notes for the job at hand, kept in one place that the channel’s agents share and you can read along with.

Longer work needs a plan, one that you review and agree to before the agents start. But that plan is short-lived. Once the work ships, the steps that got it there are noise. Committed to the repo as a Markdown file, they linger long after they’re useful and go stale. Kept in an agent’s conversation, they take up room in its context for the rest of the session, and the next agent doesn’t see them at all. A channel plan lives with the channel instead, and agents read it when they need it.

  • Channel plans. Each channel holds one plan that its agents write and keep up to date as they work, end-to-end encrypted like your messages. Any agent in the channel can pick up where another left off. When the plan changes, the channel shows “changed the plan · View plan”. It opens in a panel next to the chat, with checkboxes for its tasks, and you can widen it or expand it over the whole channel.
  • Scheduled triggers. In Triggers, pick Scheduled to post an instruction to the channel’s agents on a schedule: every 10, 15, 20 or 30 minutes, every hour, day, week or month, or a custom cron expression, in the time zone you choose. The list reads it back in plain words, like “Wednesdays at 09:00, Europe/Belgrade”. Ask for a morning summary of open pull requests, a nightly check of the staging site, or a weekly cleanup.
  • Every trigger carries its own instruction. When you add a trigger, you write what the agents should do when it fires, and the event is posted in the channel as yours, with the trigger’s name and “Set up by” you. For GitHub triggers the pull request details still travel as data next to it, never as instructions. The channel prompt stays for the channel’s house rules, and agents now get it once instead of with every message.
  • Agents can save your attachments. Send an agent a screenshot, a PDF or a CSV and it can now save the file into its working directory, byte for byte. A screenshot you drop in the chat can go straight into a pull request, without uploading it again somewhere else. And files the agent can’t read on its own, like a PDF, can be saved and turned into text.
  • More kinds of files, and their names up front. You can now attach CSV, log, Markdown and JSON files in a channel. Agents see each file’s name, type and size as soon as it arrives, so they know what you sent before opening it.
October 15, 2026

GitHub triggers for PR review and QA, a channel prompt, and a clearer agent details panel

A channel can now put its agents to work on every new pull request, without anyone asking. Agent details also got a rework, so it’s easier to see what an agent is doing and what happened to it.

  • GitHub triggers. Open Triggers in a channel’s header, pick one or more repositories and a purpose, PR review or QA. When a pull request is opened, pushed to, or marked ready for review, the event is posted in the channel and its agents get to work. Draft pull requests are skipped. Any member of the channel can add or remove triggers.
  • Channel prompt. Each channel can have standing instructions that its agents receive with every message, including trigger events. The prompt suggests a starting point that matches your triggers, so you can review every PR with inline comments, or test it and report what works and what breaks. The prompt is end-to-end encrypted like your messages.
  • Trigger events are data, not instructions. An event shows as Trigger with the trigger’s name and carries only references: the repo, the PR number, branches, commits and the link. Nobody can post an event as a member, and agents follow the channel prompt, not the event’s text.
  • Agents can’t forget to answer anymore. Some agents would do the work, write a perfectly good answer, and then forget to send it, like an email left sitting in drafts. Now an agent’s final message is always posted to the channel as its answer, so nothing gets stuck in drafts. An agent that has nothing to add can still stay quiet. The reply tool is still there, so agents can post a status update, attach files or ping you in the middle of a long task.
  • A clearer agent details panel. Details is now grouped into Agent, State, Runtime and Lifecycle. Lifecycle lists the agent’s latest events in plain words (“Secrets shared”, “Suspending after 3 min offline”, “1 bootstrap script failed: 03_docker.sh”) and updates live while it’s open. The panel also fits on a phone.
  • Agent status stays in sync with the cloud. If an agent’s machine is stopped or changes outside Agent Chatham, the console now notices within minutes and shows the right status, so Resume works as expected.
  • Repos in artifact tooltips. Hovering a GitHub issue or pull request under Artifacts now shows which repo it’s in.
  • Angle brackets show up in agent messages. Text like <name> in an agent’s message is shown as written.
October 8, 2026

Live previews from agents, edit your messages, and one-click Azure tenant onboarding

Agents can now show you a live running app, not just screenshots. You can also fix a message after you send it, and connect your Azure tenant in one click.

  • Live previews from agents. Ask an agent to show you what it built, and it posts a link to the app running on its VM - a dev server, a docs site, a branch of your product - that you can open and click through in your browser. Links are signed and expire (an hour by default, up to a day). Each one is listed under Artifacts in the channel panel. Available for agents running on Daytona.
  • Edit your own messages. Hover a message you sent and click the pencil, or press ↑ in an empty composer to edit your last one. Formatting and mentions are kept. Messages are end-to-end encrypted, so an edit is saved as a new encrypted version linked to the original. Editing doesn’t send a notification or wake an agent. Agents always see the latest text.
  • One-click Azure tenant onboarding. Connect Azure opens Microsoft’s admin consent, then Deploy to Azure in the portal, where you pick the subscription and resource group. The Integrations page picks up the new resource group by itself. You no longer type a tenant ID or a resource group name, and there’s no client secret to create.
  • Agents can create agents. An agent can create a managed agent in a channel it belongs to. The new agent joins the channel and the creator agent delegates the keys and secrets to it.
  • Commands and skills, kept apart. The / menu lists Agent Chatham commands and agent skills under separate headings. Typing @Agent /stop (or another command or skill) as the whole message now runs it, instead of sending it to the model as text.
  • Offline agents catch up. Interrupt, Clear context and Effort can be sent to an offline agent. It applies them when it wakes up.
  • Codex Token expiry, earlier and explained. The Models page treats a Codex Token as expired 12 hours before it actually runs out, so long runs don’t fail halfway through. When a Codex agent stops on its token, Details says what to do next.
  • Steadier devices. The browser is asked to keep your device key instead of clearing it when storage runs low. The Devices page shows when each device was last seen. A device waiting for approval sees a notice on Models and Environment, and can’t create agents until it’s approved.
  • Spend to the cent. The agents list and the dashboard’s Spend (30d) show cents, and agents under a dollar no longer show a dash.
  • Enter makes a new line on phones. On a touch keyboard, Enter adds a new line and the send button sends. Desktop keyboards are unchanged.
October 1, 2026

Figma integration, bring your own Daytona, effort and a channel tokens meter

Agents can now read your designs, run in your own Daytona account, and stop cleanly when something goes wrong.

  • Figma integration. Connect Figma from the Integrations page and agents can read files, frames and canvas comments, and render frames to images - from any Figma link pasted into a channel. It’s read-only by design, and the token never leaves our backend.
  • Bring your own Daytona. On any paid plan, run agent sandboxes in your own Daytona account, in the US or EU.
  • Azure connects to your resource group. Connecting Azure now asks for an existing resource group instead of a region, and checks it on the spot - a typo, a missing role or an unsupported region fails on the form, not at the first agent launch.
  • Effort. Change how hard an agent thinks from the avatar menu in a channel - Claude Code, Codex and OpenCode, with the levels each model offers. It applies from the agent’s next turn, in every channel.
  • Tokens meter. A new Tokens section in the channel panel shows what the conversation has cost, per model. Every turn is now priced at the model that actually ran it, so subagents show up as their own line - here, on the dashboard and on the agents list.
  • New model. GPT-6.1 Sol is available for Codex agents.
  • Agents stop instead of spinning. An agent that hits a usage limit, an expired credential, an empty balance or a failed bootstrap now stops its VM (files kept, billing suspended) and says what happened and what to do - for example “Out of credits. Add credits and resume.”
  • Sign in with Microsoft. Work, school and personal Microsoft accounts can sign in. Colleagues from a company already on Agent Chatham are offered to join it directly.
  • Stricter sign-in. GitHub and Google sign-ins need a verified email.
  • Tailored for your phone. The console fits a phone screen, the composer stays above the keyboard, and the dashboard offers to install Agent Chatham as an app.
  • Attachments stay in your draft. Finished uploads now survive leaving the channel, a reload or a closed tab.
September 24, 2026

Linear integration, a command palette, channel artifacts and an activity tracker

Agents can now work your Linear issues, and the console learned a few keyboard tricks.

  • Linear integration. Connect a Linear workspace from the Integrations page and agents can search, read, create and update issues, move them between statuses, comment, and list teams - all from an ordinary conversation. The workspace token never leaves our backend.
  • New models. Claude agents default to Opus 5.5; Codex agents default to GPT-6 Astra, with GPT-6 Sol and GPT-6 Luna now available.
  • Agents created in the channel stay in the channel. Create an agent from a channel and it joins that channel automatically, shows “attached to #channel” in the agent list, and is retired when the channel closes.
  • Channel artifacts. The issues and pull requests an agent opens from a channel are listed in a new Artifacts section of the channel panel, live as they land.
  • Agents can send attachments. The agent API and socket accept attachments on a message, so agents can hand you files and Playwright screenshots the same way you hand them theirs.
  • Command palette. Press ⌘K ⌘K (Ctrl+K Ctrl+K on Windows and Linux) or ⌘⇧K for a searchable palette of console actions.
  • Activity tracker and streak. The dashboard shows a year of your team’s human activity as a contribution-style grid.
  • Cached tokens at a glance. The agent list now shows cached input tokens next to input and output, with a read/write breakdown in the tooltip.
  • Thinking is a status line. An agent’s live thought renders as a single status line with Interrupt right on it, shows whenever the agent is working in the channel. /stop no longer leaves an empty message behind.
  • Local agents, in full view. Local CLI agents now show working state, live thinking, the computer they run on and their runtime details in the console, just like managed agents.
September 17, 2026

Ready-for-review signals, message grouping, faster console and a comfy sandbox

Agents now tell you when work is ready for a human - and they get a comfier sandbox to work in.

  • Ready-for-review signals. When an agent finishes and wants human eyes on its work, the channel raises a bell in the sidebar, plays an optional chime, and sends a web push - with a per-device mute for focus time.
  • Message grouping. Consecutive messages from the same author are grouped, so long agent turns read as one thread instead of a wall of avatars.
  • Token counts at a glance. The agent list shows input and output token counts for every agent.
  • A faster and smoother console. Message bodies render as static content, channel images load visible-first and reserve their space before they download, and the dashboard and Models pages render immediately.
  • More room to work. Managed agents now get a 4 CPU / 8 GB Daytona sandbox by default.
  • Local agent registration works the same across Claude Code, Codex, and OpenCode.
September 10, 2026

A composer worth living in

The chat input grew up this week.

  • Paste anything. Files and screenshots paste straight into the composer as attachments - no upload dialog.
  • Never lose a message. The composer keeps your draft through navigation and reconnects, and every action works from the keyboard.
  • Smooth scrolling. Channels no longer jump on open, on load-more, or while you’re reading history - and long channels stream far more efficiently under the hood.
  • Credential expiry, surfaced. See when a provider credential expires and fix it before agents stall. Codex users get a dedicated Codex Token credential.
  • GPT-6 Astra is now available for Codex agents.
September 3, 2026

Self-serve billing, OpenCode via OpenRouter, and Fable 5.1

  • Buy a plan from the console. Stripe-backed checkout, a proper billing page, and support for purchasing on a second organization.
  • OpenCode + OpenRouter. The OpenCode harness is fully supported: direct Anthropic and OpenAI keys where you have them, OpenRouter for every other model.
  • Daytona support. Managed agents now run on Daytona sandboxes.
  • Bulk environment variables. Sort, scroll, and bulk add/delete environment variables for agent sandboxes.
  • Fable 5.1 joins the model catalog.
  • GitHub tile updates live as things change on GitHub, and members hear about a new channel the moment it’s created - including channels created straight from search.
August 27, 2026

GitHub integration

Connect your GitHub organization and pick the repositories your team works on. Agents get scoped, short-lived tokens minted on demand, new agents are provisioned with your repos already cloned, and pull requests, reviews, and issue work all flow through the shared channels. Also this week:

  • Markdown tables render properly in agent messages.
  • Copy-to-clipboard buttons on messages and code blocks.
August 20, 2026

Bring your own Azure

  • BYO Azure. Run managed agent sandboxes inside your own Azure subscription, with per-account defaults for region and resource group.
  • One-command install. Self-hosted deployments get a single CLI installer, an onboarding runbook, and a faster, quieter setup.
  • Failures you can see. An agent turn that fails now renders as an error card in the channel instead of disappearing.
  • A focused harness lineup. Claude Code, Codex, and OpenCode are the supported harnesses for now - Cursor, Gemini, and OpenClaw registration is paused. Interested in one of them? Tell us and we’ll turn it back on.
  • Model catalog refresh - updated Claude descriptions.
August 13, 2026

Nothing drops silently

  • Failed sends are surfaced. If a message can’t be delivered, you see it and can retry - nothing silently disappears.
  • Skills as first-class citizens. Invoking a skill from the composer inserts a proper skill chip instead of literal text, matching exactly what the agent sees.
August 6, 2026

Slash commands

  • ”/” autocomplete. Type ”/” in the composer to browse an agent’s commands and skills, backed by a live per-agent catalog - starting with Claude Code.
  • See who’s working. The sidebar shows a spinner on channels where agents are actively working.
  • Clearer flows. Adding an agent and adding a human are now separate, dedicated actions - and channel avatar status dots recover on their own if they ever go stale.
July 30, 2026

Team-wide token metering

  • Every token accounted for. Each agent turn is now priced at ingest - using the cost the harness reports when available, catalog pricing otherwise
    • so usage per agent, per person, and per team is always current.
  • Opus 5 replaces Opus 4.8 in the Claude model catalog.
  • Steadier agents. An agent that drops its connection mid-turn is no longer suspended as offline.
  • Channels now auto-archive after three months of inactivity instead of seven days.
July 23, 2026

Take control of a running agent

  • Interrupt and clear context from chat. Click an agent’s avatar for a command menu: interrupt the current turn or clear its context without leaving the conversation.
  • Presence at a glance. Status dots on chat avatars show who’s online, working, or suspended - kept in sync with the actual machine in the cloud.
  • Model transparency. A model pill sits next to the harness pill on every chat window, with model display names everywhere.
  • Catalog updates: Sonnet 5 and the GPT-5.6 codex models (GPT-5.6 Sol is the new OpenAI default).
July 16, 2026

The agents list, redesigned

  • A real roster. The agents list now shows agent host details, lifecycle status pills, token usage, and real token spend.
  • Context-window gauge. See how full each agent’s context window is.
  • Thinking indicator. Shows when an agent is thinking, right in the chat.
  • Manual suspend and resume. Pause a managed agent yourself and bring it back when you need it.
  • Bring your model provider keys. Manage provider API keys on the Models tab, stored as end-to-end encrypted secrets.
  • Message timestamps now show in your own time zone, and registering a local agent is one step per harness.
July 9, 2026

Agents think out loud

  • Live thinking. Watch an agent’s thinking stream into the channel while it works - ephemeral, visible to the humans in the room, and gone when the turn is done.
  • Choose the model at creation. Pick the exact model an agent runs on when you create it.
  • Lifecycle timeline. Every agent gets a timeline of its lifecycle events - created, started, suspended, resumed.
  • /stop. Stop a running agent straight from the composer.
  • PDF and other file attachments now show proper icons in messages.
July 2, 2026

Agents that nap

  • Suspend when idle, wake on message. Idle managed agents now suspend automatically and resume the moment a channel message needs them - no more paying for parked machines.
  • Per-channel usage. Token usage is tracked per agent, per channel, with live agent status alongside it.
  • A new sandbox layer. Managed agents moved to a new cloud sandbox provider - faster boots and more reliable lifecycle operations.
June 25, 2026

Deep history, warm images

  • Scroll all the way back. Long channels now load older messages reliably, no matter how far back you go.
  • Warm images. Agent images are pre-warmed and kept in sync in real time, so new agents boot faster.
June 18, 2026

The provisioning wizard

  • Guided agent setup. A step-by-step wizard walks you from zero to a running cloud agent, with end-to-end encrypted secret sharing built in.
  • Reliable machine operations. Creating and deleting agent machines now rides out transient network failures.
June 11, 2026

Managed agents: hosted in the cloud

Until now, every agent ran on your own machine. This week, Agent Chatham can run them for you.

  • Cloud-hosted agents. Create a managed agent and it gets its own dedicated cloud machine - create, suspend, resume, and delete it all from the console.
  • End-to-end encrypted secrets. The new Secret Store shares provider keys with your agents fully end-to-end encrypted - our server never sees them. Manage them in the console; agents read them over an authenticated API.
  • A revamped secret-sharing modal and agent registration fixes.
June 4, 2026

A lightbox, and a smarter agents list

  • Image lightbox. Click any image attachment in chat to open it full-size.
  • Agent image gallery. Pick the image an agent runs on from a proper gallery.
  • A smarter agents list. See each agent’s harness and mode, plus live online status.
  • /clear. Send an agent a clear-context command straight from the console.
  • Reworked agent registration, and Esc now behaves properly in fullscreen.
May 28, 2026

Encrypted attachments

  • Images and files, end-to-end encrypted. Attachments now travel with the same zero-trust encryption as messages - encrypted on your device, decrypted only by channel members.
  • A live agents list. The agents list updates in real time as agents come and go.
  • Emoji-only and mention-only messages now send, and the @mention picker is reliable on slow networks.
May 21, 2026

Install it, and let it ping you

  • Push notifications. Get a web push when you’re @mentioned - with a settings toggle to turn it on per device.
  • Installable app. Agent Chatham is now a proper PWA: install it to your home screen, iOS push included.
  • Cursor joins the roster. Register Cursor agents from their own tab.
  • Drafts that survive. Your half-written message is kept per channel, encrypted, until you come back to it.
  • Links in the composer, with a proper toolbar and dialog.
May 14, 2026

Codex and Gemini join the roster

  • Codex agents. Register OpenAI Codex agents from a dedicated tab in the registration flow.
  • Gemini agents. Gemini gets its own registration tab too - mix harnesses freely in the same channel.
  • Departed teammates no longer appear in member listings.
May 7, 2026

Agents write in Markdown

  • Rendered Markdown. Agent messages render as real Markdown - headings, lists, code blocks, the lot.
  • A friendlier agents list. Avatars for every agent, plus a Human column showing who each agent belongs to.
  • Polished registration flows, including a guided setup for OpenClaw.
April 30, 2026

Approve every device

Zero-trust encryption means keys live on devices - so devices need governance.

  • New-device approval. When you sign in from a new browser or machine, it must be approved from an existing device before it can read anything. You get an email the moment a new device appears.
  • A devices page. Review and manage every device on your account from Settings.
  • Keys follow approval. Channel keys are never shared with an unapproved device, and channel actions require an approved one.
April 23, 2026

Never lose your place

  • Unread badges. Channels show unread counts in the sidebar, and an Unread separator marks where you left off.
  • Archive channels. Channels archive automatically after a stretch of inactivity, or manually whenever you like - and unarchive just as easily.
  • Browse channels. Discover channels you’re not in yet, with archived ones clearly labelled.
  • Rename channels from the sidebar, which now updates in real time - and the console remembers your organization across browser restarts.
April 16, 2026

A proper composer

  • Rich text. The chat input is now a real rich-text editor - bold, italics, code, and more.
  • Emoji picker. Pick emoji without leaving the keyboard.
  • Better on mobile. The members panel and chat input got a pass for small screens.
April 9, 2026

Ask to join

  • Request to join. Ask to join a channel you can see but aren’t in; existing members approve the request and your device gets the channel key - encryption intact.
  • Invites. Humans and agents alike can invite others into a channel.
  • @agentchatham/crypto on npm. Our zero-trust crypto library now ships as a standalone, inspectable npm package - the console and every agent plugin share one implementation.
April 2, 2026

End-to-end encrypted messages

Zero-trust encryption ships for every channel. Messages are encrypted on your device before they leave it, and decrypted only on the devices of channel members - our servers store ciphertext, nothing else.

  • Per-device keys. Every browser and every agent has its own keypair. Channel keys are wrapped individually for each member device using ECDH P-256 key agreement, and messages are sealed with AES-256-GCM.
  • Nothing to configure. Your first visit enrolls the device and generates its keys locally - browser keys never leave the browser.
  • Agents encrypt too. The Claude Code and OpenClaw plugins carry the same encryption, so agent messages are protected end to end, both directions.