Standcast

Public beta

Standcast is in public beta. Expect rough edges and changes to behaviour, configuration and the API before 1.0. Report problems in GitHub issues or by email (see Privacy).

Standcast runs your team's daily standup. Each person writes their daily with their AI agent, starting from a draft of what they did on GitHub. Standcast collects the dailies and posts one digest per team to Slack at the time you choose.

What it is

Standcast is two things that work together:

• A GitHub App that your organization admin installs. It reads the team setup from a standcast.yml file, receives the dailies and posts the team digest to Slack.
• A skill, a CLI and an MCP server that each person uses with their AI agent (Claude Code, Cursor, Codex, Claude Desktop, VS Code). The agent fetches a draft of yesterday's work, the person fixes it, writes today's plan and blockers, and confirms before anything is sent.

It solves three problems that show up once a team writes dailies with AI:

• Loose dailies. Every message comes in a different format, in a different place.
• No consolidated view. The manager reads scattered messages and has no single picture of the team.
• Nobody orchestrates. No one tracks who sent the daily, and blockers sit until someone notices.

Who it is for

• Members write their own daily through their AI agent. Start with Getting started.
• Managers get the team digest in Slack and can read their team's dailies, digests and blockers from their own AI agent. Start with MCP server.
• Admins install the GitHub App, connect Slack and write standcast.yml. Start with Install.
• Integrators read organization data with an API key and receive webhooks. Start with the API overview.

How it works

  Member's AI agent                    Standcast                         Slack
 ┌──────────────────────┐   draft   ┌────────────────────────┐      ┌───────────────────┐
 │ /daily skill         │ ◀──────── │ API                    │      │ #daily-backend    │
 │ standcast CLI        │           │  · reads GitHub        │      │                   │
 │ standcast mcp        │ ────────▶ │  · stores the daily    │ ───▶ │ one digest per    │
 │ (person confirms)    │  submit   │ Worker                 │      │ team per day,     │
 └──────────────────────┘           │  · digest at digest_at │      │ late dailies in   │
                                    │  · late daily → thread │      │ the thread        │
                                    └────────────────────────┘      └───────────────────┘
                                                 │
                                                 └──▶ API keys and webhooks for your systems

1. The person runs /daily in Claude Code, or asks any MCP client for their daily.
2. Standcast builds a draft of "yesterday" from the person's GitHub activity since their last daily: pull requests opened and merged, reviews and commits.
3. The person edits yesterday, writes today and any blockers, and confirms. Nothing is sent before that.
4. At digest_at, in the team's time zone, Standcast posts one message to the team channel: blockers on top, then each person's daily, then who did not send it.
5. A daily sent after the digest goes as a reply in the digest thread.

Around it, Standcast orchestrates the day:

• at remind_at, people who have not sent the daily get one direct message;
• each new blocker goes by direct message to whoever can help: the pull request's reviewers, its CODEOWNERS, or the team managers;
• suggested blockers, such as a pull request waiting for review or with failing CI, appear only to you, in your own draft;
• every change to standcast.yml is validated by a GitHub check run.

Linear cards show up in the draft and in blocker routing. Coming soon: Google Chat and Jira.

Read How it works for the full cycle of a day.

What Standcast does not do

• It is not a surveillance tool. The person always has the final word on their own daily. Nothing is sent on someone's behalf without their explicit confirmation.
• It does not measure productivity. There are no rankings, scores, commit counts per person or per-person aggregates, in the digest, the API or the MCP tools.
• It does not write the daily for you. The draft is a starting point. Today's plan and the blockers are written by the person.
• It does not read your code or your Slack messages. See Privacy.
• It does not send your data to an AI model. No LLM processes your data on the Standcast servers. Your own agent runs on your machine, with your own account.

Start here

You are	Read
An admin setting up your organization	Install, then the standcast.yml reference
A member writing your first daily	Getting started
A Claude Code user	Claude Code skill
A Cursor, Codex, Claude Desktop or VS Code user	MCP server
A manager	MCP server and How it works
An integrator	API overview and Webhooks

New to the terms? See Concepts.

For AI agents

Standcast publishes its documentation in a format built for AI agents. llms.txt lists every page, llms-full.txt holds the whole documentation in one file, and every page is also available as Markdown by adding .md to its address. The Portuguese files are at /pt-BR/llms.txt and /pt-BR/llms-full.txt.