Integrations are the foundation of the platform. Every user story Arga generates — for a PR, a URL, or a sandbox run — is written from the live context it pulls out of the tools your team already runs on. Source control tells the platform what changed; chat and issue trackers tell it what users care about and what’s been broken; observability tools tell it what production actually looks like. Without integrations, the platform is guessing. With them, it’s writing the user stories you would have written if you had time.Documentation Index
Fetch the complete documentation index at: https://docs.argalabs.com/llms.txt
Use this file to discover all available pages before exploring further.
Connecting integrations
Go to app.argalabs.com/integrations and click Connect on the tools you use. Each integration uses OAuth — you’ll be redirected to the provider to authorize access, then returned to Arga automatically. The more sources you connect, the sharper and more grounded the auto-generated user stories become.Context sources
Arga currently reads context from the following tools:| Category | Supported tools |
|---|---|
| Source control | GitHub |
| Communication | Slack, Discord |
| Issue tracking | Jira, Linear, GitHub Issues |
| Error monitoring | Sentry |
| Observability | Grafana, PostHog |
Integration settings
Some integrations support user-configurable settings that control automatic behaviors during validation runs.Linear
When you connect Linear, Arga automatically creates issues in your default Linear team for failed validation steps. You can toggle this off from the Integrations page by unchecking Auto-create issues for failed runs on the Linear card. You can also update this setting via the API:Sandbox twins
Context integrations are how the platform understands your stack. Sandbox twins are how it stands in for external services so generated user stories can run safely against a real-looking surface — no production calls, no flaky third-party dependencies. Arga currently supports these twins in sandbox environments. Twins are split into two categories: frontend twins, which expose a browsable sandbox surface in the Arga UI during a run, and backend twins, which intercept outbound API traffic only. Frontend twins| Twin | Purpose |
|---|---|
discord | Stand in for Discord API v10 calls (discord.com, api.discord.com, discordapp.com) during sandbox validation. Supports the full Discord API v10 surface including guilds, channels, messages (with multipart/form-data file attachments), threads, members, roles, webhooks, emojis, stickers, scheduled events, stage instances, and auto-moderation. |
dropbox | Stand in for Dropbox API v2 calls (api.dropboxapi.com, content.dropboxapi.com, notify.dropboxapi.com) during sandbox validation. Supports the full Dropbox API v2 surface including files, sharing, file requests, file properties, team management, and admin controls. |
github | Stand in for GitHub REST API calls (api.github.com, github.com) during sandbox validation. Supports repositories, pull requests, issues, branches, commits, check runs, search, organizations, users, and webhook delivery with signed payloads. |
google_calendar | Stand in for Google Calendar API calls (www.googleapis.com, calendar.googleapis.com) during sandbox validation. Supports calendars, events, ACL rules, settings, free/busy queries, and webhook deliveries. |
google_drive | Stand in for Google Drive API calls during sandbox validation. |
notion | Stand in for Notion API calls (api.notion.com, notion.so) during sandbox validation. Supports pages, databases, blocks, users, and search. |
slack | Stand in for Slack Web API calls during sandbox validation. Supports channels, messages, conversations, files (including external uploads that automatically create channel messages), and the OAuth 2.0 authorization code grant flow for testing “Add to Slack” install flows. |
stripe | Stand in for Stripe API calls during sandbox validation. |
| Twin | Purpose |
|---|---|
box | Stand in for Box API calls (api.box.com, upload.box.com) during sandbox validation. |
jira | Stand in for Jira Cloud REST API v3 calls (*.atlassian.net) during sandbox validation. Supports issues, projects, comments, worklogs, attachments, issue links and properties, votes, watchers, JQL search, users, components, versions, remote links, filters, dashboards, groups, fields, metadata (priorities, resolutions, statuses, status categories, issue types), webhooks, and Agile boards and sprints. Comments and descriptions accept both Atlassian Document Format and plain strings. Includes the OAuth 2.0 authorization code grant flow with scope enforcement on every authenticated route. |
unified | Stand in for the Unified API when your app talks to api.unified.to instead of directly to the upstream provider. |
unstructured | Stand in for Unstructured API calls (api.unstructuredapp.io, platform.unstructuredapp.io) during sandbox validation. Supports document partitioning, workflow management, source and destination CRUD, and job execution. |
What Ask can do with each integration
When you use Ask to investigate a bug or search for context, the chat agent has access to specific capabilities for each connected integration. The more integrations you connect, the deeper Ask can dig.Sentry
- Search issues — find error issues by message, exception type, or culprit function
- List events for an issue — retrieve individual error events with stack traces, breadcrumbs, and release information
- Get full event details — drill into a single event for the innermost exception frames and the breadcrumbs leading up to the error
Grafana
- Get alerts — retrieve active or recent alerts, filtered by state (alerting, ok, pending, no data)
- Search dashboards — find dashboards by name
- Query Prometheus metrics — run PromQL range queries through a Grafana datasource to fetch raw metric series (latency, error rate, throughput) over a time window
- Fetch annotations — retrieve deploy markers, incident notes, and manual events to correlate incidents with deploys
PostHog
- Search errors — find exception events by error message or type
- Run HogQL queries — execute custom SQL queries against PostHog events for ad-hoc analysis
- Get session recording metadata — fetch duration, event count, and a deep link to replay a specific session recording
Discord
- Search messages — search messages across a guild by keyword, optionally scoped to a channel
- Get channel history — fetch the most recent messages in a channel for chronological context around an incident discussion
Other integrations
| Integration | Capabilities |
|---|---|
| GitHub | Search PRs and issues, get PR details (changed files, diff summary, review comments), get commit details |
| Slack | Search messages and threads |
| Jira | Search issues |
| Linear | Search issues, get issue details |
| Notion | Search pages and databases |
Supported observability and logging tools
Supported tools
Supported tools
- Grafana
- Sentry
- PostHog
- Jira
- Linear
- GitHub Issues
Coming soon
Coming soon
- Datadog
- Prometheus
- AWS CloudWatch
- AWS X-Ray
- Loki
- Tempo
- ELK (Elasticsearch, Logstash, Kibana)
- Jaeger
- Splunk
- Honeycomb
Next steps
Run your first validation
Get up and running in under a minute from the terminal.
Talk to our team
On a Team or Enterprise plan? Schedule a guided onboarding session to connect your tools together.