Svix Changelog September 2026

SDK v2 is stable

Last month we shipped the release candidates, and now v2 of the SDKs and CLI is out for real, covering JavaScript/TypeScript, Python, Go, Rust, Java, Kotlin, C#, Ruby, and PHP. v2 cleans up long-standing naming quirks: update operations are now upsert (they were always create-or-update), filterTypes is now eventTypes, deprecated operations are gone, and with_content now defaults to false so large payloads aren't mirrored back unless you ask for them.

See the v2.0.0 release notes for the full list of changes and per-language migration notes.

Svix CLI wizard

The Svix CLI has a new svix wizard quickstart command: a guided, interactive walkthrough of your first Svix integration. It logs you in, creates an application and endpoint, and sends your first message, right from the terminal.

It can also install the Svix agent skills for you, so you can hand the rest of the integration to your coding agent with all the right context already in place.

App Portal MCP server

We launched the App Portal MCP server, which lets your customers connect their coding agent to the App Portal and debug their webhook integration without leaving their editor. The agent can read endpoints, failed delivery attempts, payloads, and responses, and even resend messages or recover an endpoint.

Like the App Portal itself, it's white-labeled, so it shows up to your customers' agents under your brand. Read more in the announcement post.

Multi-organization membership

You can now be a member of multiple Svix organizations with a single account. Accepting an invitation no longer moves you out of your existing organization; instead you get an organization switcher in the dashboard and can move between them freely.

This is especially useful for agencies and consultants working across customers, and for companies that keep separate organizations for separate products.

More advanced endpoint types are now public

The full set of sink types, including AWS SQS, SNS, EventBridge, Google Cloud Pub/Sub, RabbitMQ, BigQuery, ClickHouse, Snowflake, and Redshift, is now publicly available, both as advanced endpoints and as Stream sinks. For Stream, they're also available in the API and SDKs as of v2.3.0.

Postgres advanced endpoints

Speaking of advanced endpoints, there's a new one: Postgres. Point it at your database and Svix delivers messages straight into a table, no consumer code required. It joins the existing storage and data warehouse options as another way to consume webhooks without running a webhook handler.

Channels support for stream sinks and advanced endpoints

Stream sinks and advanced endpoints now support channels, so you can filter which messages each sink receives the same way you already can with regular endpoints. Set up to 10 channels on a sink and it will only receive messages tagged with them.

Delivery attempt tracking for sinks

Stream sinks and advanced endpoints now track individual delivery attempts, so you can see exactly what was attempted, when, and how it went, the same visibility you've always had for regular webhook endpoints. This makes debugging a misbehaving sink much less of a guessing game.

Choose which advanced endpoint types are enabled

You can now enable or disable specific advanced endpoint types per environment, instead of the previous all-or-nothing toggle. Want your users to create polling endpoints and S3 buckets but not data warehouse connections? Pick exactly the types you want to offer from the environment settings.

Polling endpoints v2 rollout complete

The v2 polling endpoints API we announced last month is now fully rolled out and is what all new polling endpoints use. v2 uses explicit offset commits, so consumers can process messages at their own pace and pick up exactly where they left off, with uncommitted messages redelivered automatically.

Cancel pending messages before a cutoff

There's a new endpoint-level API to cancel all pending deliveries created before a given timestamp. If an endpoint has accumulated a backlog of messages that no longer make sense to deliver, say after an extended receiver outage, you can now draw a line and only deliver what came after it, instead of flooding the receiver with stale events.

Programmatic access to organization members

You can now list your organization's members via the management API, including each member's role and their effective access in each environment. Useful for access audits and keeping your internal systems in sync without clicking through the dashboard.

Bulk purge message contents

The new bulk expunge API deletes the payloads of a specific set of messages in a single call. It's the tool for the "we accidentally sent sensitive data" moment: pass the message IDs and the payloads are gone, while the delivery metadata stays intact for your records. It complements the existing expunge-all-contents endpoint, which wipes payloads for an entire application.

It's available in the SDKs as of v2.1.0, and in Svix Server OSS as of v1.101.0.

Closing words

As always, please let us know if you have any feedback on the product, or anything else we can add or build to make your lives easier. A lot of the Svix product came from customer feedback, and we are sure ${next_big_thing} will come from feedback as well. No feedback is too small!

You can contact us directly or join the discussion on our community Slack.

For the latest updates, make sure to follow us on X, Github, RSS, the Svix newsletter, and join the discussion on our community Slack.