AI Agents — Release Notes
v1.1.0 — TBD
New Features
Custom Agent Builder
You can now create your own AI agents from the admin panel. Go to Admin → Agent Builder → New Agent to get started. Each custom agent you create works exactly like a built-in agent — it appears in the workspace, respects plan entitlements, and can be included in pipelines.
Agents you publish are immediately available to eligible subscribers. Set them to Draft to keep them hidden while you're building.
Agent Overrides
You can now change the display name or description of any built-in agent, or hide it from users entirely, without touching any code. Hidden agents are still usable in pipelines — only the user-facing workspace card is hidden. Find overrides at Admin → Agent Builder under the Built-in Agents section.
Pipeline Builder
Chain multiple agents together into an automated multi-step workflow. Go to Admin → Pipelines → New Pipeline to build one visually. Each step uses the output of the previous step as input, and you control the mapping using simple token syntax — for example, {{input.keyword}} or {{step1.output.content}}.
Pipelines run sequentially in the background. Users fill in the pipeline's starting inputs and the rest runs automatically.
Plan Entitlements Editor
Agent run limits and feature access are now configurable from the admin panel at Admin → Plans. Select a plan, add or edit entitlements, and the changes take effect for new billing periods without any code changes.
v1.0.0 — April 27, 2026
First public release of the AiAgentsCore module for ChargePanda.
Not included with ChargePanda
AiAgentsCore is a separate, paid module available from the ChargePanda Marketplace.
What's included
AiAgentsCore is the engine that powers AI-agent features on ChargePanda. It is a developer-facing foundation module — you do not use it directly; instead, other modules (such as SEO Agents) are built on top of it.
If you purchased SEO Agents, AiAgentsCore is included in the download and installs automatically alongside it.
Features
| Feature | Detail |
|---|---|
| Multi-provider AI | Supports OpenAI, Anthropic (Claude), Google Gemini, Groq, Mistral, and self-hosted Ollama |
| Plan-gated access | Agents are locked or unlocked per subscription plan using ChargePanda's entitlement system |
| Monthly run limits | Enforced per billing period, defined in plan settings in the admin panel |
| Background execution | Runs dispatch to a dedicated queue — customers are never left waiting on a loading screen |
| Live status polling | The UI updates every two seconds until a run completes |
| Cost tracking | Optional per-token cost calculation — set your rates in admin settings |
| Completion emails | Optional email to the customer when a long-running agent finishes |
| Admin playground | Test any agent directly from the admin panel without affecting billing |
Requirements
- ChargePanda v1.3.0 or later
- A queue worker running on the
agentsqueue - An API key for at least one supported LLM provider