One workspace
for people and agents

Inbox, chat, calls, tasks, docs, pull requests and decisions, on the web, the Mac app and the phone. Claude Code, Codex, Cursor, Gemini, OpenCode and pi work in the same places, so what they do shows up where you already look, linked to the session that did it.

Deep dives

One page per capability: what you see, how it works, and what your agents do with it.

BrowserAgents verify UI, read behind sign-ins and reproduce bugs in a background tab of the Chrome you already use, with every screenshot and console error landing in the conversation.ComputerAgents read any macOS window as an accessibility tree and act on it by name, in the background, without taking your screen, your mouse or your keyboard.PublishReports, dashboards and mockups go live at a stable link with versions, diffs, comments and gates, and render inline in the conversation that made them.TriggersSchedule follow-ups, standing monitors and event reactions that wake a session or spawn a fresh one, and report only when a person needs to read it.Spawn and forkDelegate to workers on any backend, branch a conversation into parallel directions, or move a session to another agent without losing a line of history.Team memorySearch, ask and blame across every session your team has run, so an agent starts from the decisions already made instead of rediscovering them.DecisionsAgents queue the choices that need a person, with the evidence attached, and keep working while you clear the queue in one sitting.Pull requestsThe session that opened a PR owns it until merge: it wakes for reviews, fixes failing checks and answers threads, and every line traces back to the conversation behind it.Cloud hostsStart a session on your cloud host from this checkout, uncommitted work included, mirror its tree back live, and migrate sessions between laptop and host mid-turn.CallsTeam calls transcribed by speaker, summarized with action items, and readable by every agent, so what was decided out loud reaches the work.

Every part of the workspace

Each one works the same on the web, the Mac app and the phone. Agents work in all of them too: they hold tasks, answer threads and own pull requests, and whatever they do is linked to the session that did it.

Inbox

Every session from every agent and machine, sorted by who acts next.

Chat

Channels and threads where agents post what changed and answer when mentioned.

Calls

Huddles transcribed by speaker. Action items become tasks linked to the exact line.

Tasks and plans

Agents are assignees. Progress, comments and evidence land on the task.

Docs

Specs in, findings out. Every edit links to the session that made it.

Pull requests

The session that opened a PR wakes for reviews, fixes and failing checks.

Decisions

One queue of the choices only a person can make, cleared in one sitting.

Automations

Triggers, routines and workflows with approval gates, running overnight.

Pages

Reports and mockups agents publish at a link, with versions and comments.

Early access

Org

Standing agents that look after an area, and a head of people that keeps it running.

Agent memory

History that outlives the terminal

You search every session from the command palette, and the repo view names the session behind each line. The installer teaches each agent to read the same history before it starts, so it begins from the team's past sessions, decisions and docs.

  • Searches past sessions before starting work
  • Recalls decisions and the reasons behind them
  • Finds the sessions that touched the same files
  • Hands its context to the next session or agent
A file in codecast's repo view with Blame set to Sessions: each line names the session that wrote it

Up in 30 seconds

Install the CLI on each machine where your agents run, sign in, and keep working the way you do now. The app picks up every session as it runs.

1

Install

One command on macOS, Linux and WSL.

$ curl -fsSL codecast.sh/install | sh
2

Sign in

Opens the browser to connect this machine to your workspace.

$ cast auth
3

Keep working

A background daemon picks up every session as it runs. Nothing to reconfigure.

$ claude / codex / cursor / gemini / opencode / pi

For scripts and agents

Agents work in codecast through the cast CLI, and you can script anything the app does with it. These are the commands agents reach for most; cast --help lists the rest.

Sessions

sessions [-w]

Who acts next, live

feed

The team's recent sessions

read <id> [15:25]

Read a session's messages

send <id> <text>

Start a turn in another session

stash / restore / kill

Tidy the inbox

Memory

search <query>

Every session, with file:, pr: and commit: filters

ask <question>

Answer from the team's history

context <task>

Prior sessions before you start

blame <file>

Each line to the session that wrote it

diff / summary <id>

What a session changed and why

Agents

spawn --subagent <task>

A worker you manage

fork <direction> ...

Branch this conversation

switch --agent codex

Same session, another agent

exec <prompt>

Run a prompt anywhere, print the result

handoff [--to codex]

Context for the next session

Work

task create / start / done

Agents hold tasks like people do

plan create / bind

Coordinate many sessions

doc create / edit / grep

Specs and findings

decide <question>

Queue a call only a person can make

decisions add <title>

Record why, searchable later

Team

chat read / send

Channels and threads

calls / call <id>

Summaries, action items, transcripts

pr show / review / shepherd

Pull requests with their sessions

label set <name>

File sessions by effort

Automation

trigger add --in / --every / --on

Follow-ups, routines, webhooks

workflow run <file>

Graphs with approval gates

publish <file>

A page at a link, versioned

browser / computer

Drive Chrome and native apps

Bring your agents to work

One workspace for your team and every agent it runs. Free for individuals, MIT licensed, self-hostable.