> For the complete documentation index, see [llms.txt](https://docs.taskade.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.taskade.com/changelog/changelog-2026/april-2026/april-30-2026.md).

# April 30, 2026

{% hint style="info" %}
**Developer highlights.** This section tracks API, MCP, SDK, automation, and Taskade Genesis platform changes for developers. For the full product changelog see [taskade.com/changelog](https://www.taskade.com/changelog); for release newsletters see [taskade.com/blog/updates](https://www.taskade.com/blog/updates).
{% endhint %}

## What's New:

### Additional Automation Actions

The following actions shipped alongside the larger batch documented in the [April 28 entry](/changelog/changelog-2026/april-2026/april-28-2026.md) and were not listed in the main table:

| Action              | Description                                                                                                                                            |
| ------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `utils.jsonExtract` | Extract one or more values from a JSON string using a JSONPath expression (powered by `jsonpath-plus`). Supports bracket notation and array wildcards. |
| `setProjectTitle`   | Rename a project by setting its title to any string value. Completes the project-management action set (pin, tag, date range, and now title).          |

**Model selection expanded to remaining AI actions:** `promptAi`, `respondAi`, and `generateAi` now accept an optional `modelName` parameter, bringing them to parity with `askAi` and `categorizeAi`. All AI automation actions now expose an AI credits consumed value as an output variable, allowing flows to track credit burn per step.

See [Actions & Triggers](/automations/automation/actions.md) for the full reference.

### New LLM Models: Qwen 3.6 Max, Kimi K2.6, DeepSeek V4

Four additional models are now available via Vercel AI Gateway (initially gated to non-production tiers; expected to graduate to full availability):

| Model                | Provider                                 | Credit cost (launch, per 1M tokens) |
| -------------------- | ---------------------------------------- | ----------------------------------- |
| Qwen 3.6 Max Preview | Alibaba (`alibaba/qwen-3.6-max-preview`) | 25                                  |
| Kimi K2.6            | Moonshot AI (`moonshotai/kimi-k2.6`)     | 15                                  |
| DeepSeek V4 Pro      | DeepSeek (`deepseek/deepseek-v4-pro`)    | 15                                  |
| DeepSeek V4 Flash    | DeepSeek (`deepseek/deepseek-v4-flash`)  | 5                                   |

The `moonshotai` and `deepseek` provider prefixes are new; both are wired through the existing Vercel AI Gateway LLM routing layer alongside the models added on April 17.

### Genesis Automation Recipes — Trigger Intent Routing

The Genesis AI builder (EVE) now prefers **Mailhook**, **Schedule**, and **Webhook** triggers when generating automation recipes whose stated intent matches those patterns. Previously, EVE would default to a manual trigger regardless of the described workflow. This improves the quality of generated automations for event-driven and scheduled use cases.

See [Automation integrations](/automations/automation/integrations.md) for connector and trigger context.

### Agent Skills Support (shipped April 24)

Agents can now have typed **skills** — structured command definitions that shape how the agent reasons about available capabilities. Skills are parsed from agent command configurations and made available to the agent as a read-only set of capabilities at runtime.

* Agent command schemas are converted into skill definitions automatically.
* Skills are surfaced in the bash sandbox so EVE and space agents can introspect available operations at runtime.
* Skills are hydrated lazily at runtime, so they can be updated without redeploying the agent.

### ARRAY\_STRING Input Type for Automation Inputs (shipped April 28)

`ARRAY_STRING` is now a supported input type for automation inputs, allowing skill and flow inputs to accept typed string arrays. This enables richer parameter schemas for skills and automation actions that need to accept multiple string values in a single field.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.taskade.com/changelog/changelog-2026/april-2026/april-30-2026.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.