# Ray for technical teams

> Your work is tracked. Written for technical teams, with the tools they use and the questions they ask first.

Source: https://heyray.io/for-technical-teams

## The week this fixes

- Somebody describes a real bug in a channel, three people react, and nobody files it.
- Standup is thirty minutes of reading tickets aloud that everyone could have read.
- Half the sprint's blockers are one person waiting on another person's review, and nobody notices for two days.
- The incident is handled in the thread and written up, badly, on Friday.

## What that looks like in a channel

**"The export breaks when a CSV has an empty first column."**

Ray writes it up as an issue with the reproduction steps from the thread, the right labels and the right project, and shows you the card before it files.

Tools: Linear, Jira, GitHub

**"What is blocking the release?"**

Ray checks the live state of the issues, the open pull requests and the review queue, then answers in the thread with what is actually waiting and on whom.

Tools: Linear, GitHub, GitLab

**"Did anyone reply to that Sentry alert?"**

Ray answers from what it did and what it can see, and says plainly when it does not know rather than guessing.

Tools: Sentry, PagerDuty

**"Digest of open bugs every weekday at nine."**

Ray posts it, and you change the time or the contents by asking in a sentence rather than editing a config.

Tools: Linear, Jira

**"Comment on the PR that we are waiting on the migration."**

Ray drafts the comment and shows you the approval card with the exact text. It posts once somebody taps Approve.

Tools: GitHub, GitLab, Bitbucket

## The tools this touches

Linear, Jira, GitHub, GitLab, Sentry, PagerDuty, Notion, Confluence, Figma, Vercel, PostHog, Productboard

Every one has its own page saying what Ray reads there and what waits for an approval. The full set is at https://heyray.io/integrations.md

## What it costs

Free covers 25 tasks a month plus a one-time 50-task starter grant that never expires. Pro is one flat price for your team's size: $49 a month for 1 to 3 people, $99 a month for 4 to 10 people, $199 a month for 11 to 25 people. No seats, and using Ray more never costs more. Failed work is never counted and never billed.

## What technical teams ask first

**Can it merge, deploy or force-push?**

No. Ray's write set is curated per tool, and the destructive operations are not in it. What it can do is listed on each integration's own page, including what it refuses, and every write it can do still stops at an approval card.

**How does it fail?**

Honestly, in the channel that asked. Ray says what went wrong rather than inventing a result, it never silently retries something destructive, and failed work is never counted or billed.

**Can we turn approvals off for the boring ones?**

No, and that is deliberate. Reads run free, so asking Ray about state costs nothing and waits for nobody. Every mutation waits for a human. It is built into the code around the tools, not a setting in a dashboard.

**Does it train on our code or our messages?**

No. Your messages are never used to train models, and analytics carry metadata only, never message content.

**What model is it?**

Ray runs on frontier models through their APIs, and the same model on every plan. There is no cheaper model on the cheaper tier.