NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
Packagist, #1798 most downloaded on Packagist
A unified layer for building AI-native applications in PHP.
Latest stable
4.1.22
Last release 16 days ago
05 Nov 2026
Ships on a steady schedule
a new release about every 1 weeks
Nearly every release is documented
notes for 55 of the last 60 stable releases
Nothing withdrawn
no release was ever pulled
2 years old
486 releases · first in 2025
Drag across the ruler to compare, or tap a release to open it.
Make zai url customizable - PR #589 by @rafaelchavesfreitas
Make zai url customizable - PR #589 by @rafaelchavesfreitas
Deprecate the InspectorObserver, use the new observer from the Inspector PHP package - Inspector/Neuron/InspectorObserver
Deprecate the InspectorObserver, use the new observer from the Inspector PHP package - Inspector/Neuron/InspectorObserver
https://docs.inspector.dev/guides/neuron-ai
Fix MeilisearchVectorStore - PR #590 by @bytestream
Your coding agent can read these notes before it upgrades. Set up the MCP server →
Fix MeilisearchVectorStore - PR #590 by @bytestream
Nothing published for this version
Collect Cached Input Tokens, Reasoning Tokens across providers - PR #572 by @mattmilesi
Collect Cached Input Tokens, Reasoning Tokens across providers - PR #572 by @mattmilesi
Support nullable for Enum validation rule in structured output - PR #587 by @bytestream
Fix Gemini structured output collision with DTO properties named "$type" - by @ilvalerione
Nothing published for this version
Fix typo in README - by @meichuanyi
Make the LogObserver easy to extends - PR #582 by @bytestream
Nothing published for this version
Improve streaming performance for Guzzle and Amp stream adapters
Improve streaming performance for Guzzle and Amp stream adapters
Improve streaming performance for Guzzle and Amp stream adapters
Improve streaming performance for Guzzle and Amp stream adapters
Nothing published for this version
Fix tracking request ID on MCP tool call
Fix tracking request ID on MCP tool call
Fixed: Null pointer dereference in HttpException::networkError()
HttpException::networkError()SseHttpTransportSseHttpTransport::buildHeaderString()HttpRequestreadLine() in stream adaptersSseHttpTransportAmpHttpClient multipart pathcatch block in AmpStream::readLine()Optimize and secure RAG components (Vector similarity, text splitters)
Optimize and secure RAG components (Vector similarity, text splitters)
Improve MCP components lifecycle - PR #575 by @tw2066
Fix tool array in ToolCallMessage
Fix tool array in ToolCallMessage
More efficient tool search with fuzzy and tokenized keywords
More efficient tool search with fuzzy and tokenized keywords
Tool search inject system prompt to ensure the agent behave as expected
Tool search inject system prompt to ensure the agent behave as expected
By default every time the provider is invoked all tools are loaded and transmitted to the backend LLM. Complex production agents connected to email, c
By default every time the provider is invoked all tools are loaded and transmitted to the backend LLM. Complex production agents connected to email, calendar, drive, CRM, and and multiple MCP servers can easily reach hundreds of tools, each carrying its name, description, parameter schema, and usage hints.
This release introduces the global middleware ToolSearchMiddleware that you can attach to your agent:
class MyAgent extends Agennt
{
...
/**
* Define the global middleware.
*/
protected function globalMiddleware(): array
{
return [
new ToolSearchMiddleware([
MyCustomTool::make(),
...CalculatorToolkit::make()->tools()
...MCPConnector::make([...])->tools()
]),
];
}
/**
* Provide core tools to the agent.
*/
protected function tools(): array
{
return [
// A list of core tools that the model always has available
TavilySearchTool::make(...),
];
}
}
Here is the full documentation: https://docs.neuron-ai.dev/agent/middleware#tool-search
Nothing published for this version
Added new Gemini models with support for structured output with tools
Added new Gemini models with support for structured output with tools
support reasoning content in Deepseek and Alibaba providers - PR #573 - by @tw2066
support reasoning content in Deepseek and Alibaba providers - PR #573 - by @tw2066
fix Gemini function call with custom tools - PR #574 by @marceloandrader
fix Gemini function call with custom tools - PR #574 by @marceloandrader
Custom Tool Call Tracking for Max Runs - PR #566 by @tryvin
Custom Tool Call Tracking for Max Runs - PR #566 by @tryvin
class MyCustomTool extends Tool implements HasRunKey
{
...
public function getRunKey(): string
{
return $this->name . ':' . json_encode($this->inputs);
}
}
Nothing published for this version
MariaDBVectorStore: https://docs.neuron-ai.dev/rag/vector-store#mariadb
Add Alibaba dash scope model platform provider - PR #560 by @tw2066
Add Alibaba dash scope model platform provider - PR #560 by @tw2066
Improve the Workflow Executor design
Improve the Workflow Executor design
Nothing published for this version
This release introduce the ability of the workflow component to run parallel branches: https://docs.neuron-ai.dev/workflow/loops-and-branches#parallel
This release introduce the ability of the workflow component to run parallel branches: https://docs.neuron-ai.dev/workflow/loops-and-branches#parallel-branches
When you want to call the execution of multiple branches in parallel, you need to return the special event ParallelEvent from your node.
use NeuronAI\Workflow\Events\ParallelEvent;
class DocumentProcessing extends Node
{
public function __invoke(StartEvent $event, WorkflowState $state): ParallelEvent
{
// Node logic here...
// Finally return a ParallelEvent
return new ParallelEvent([
'text' => new TextProcessEvent(),
'image' => new ImageProcessEvent(),
]);
}
}
Several improvements and fixes to the AWS Bedrock provider - PR #562 by @vitaliiivanovspryker
Several improvements and fixes to the AWS Bedrock provider - PR #562 by @vitaliiivanovspryker
Make Http clients implementations ready for extension - Fix #557 #558
Make Http clients implementations ready for extension - Fix #557 #558
Fix Typesense vector store - #552
Fix Tavily search tool - PR #561 by @AlexTr
Fix Tavily search tool - PR #561 by @AlexTr
Nothing published for this version
Fix tool content deserialization in chat history - Fix #550 Add options argument to GuzzleHttpClient - Fix #551
Fix tool content deserialization in chat history - Fix #550 Add options argument to GuzzleHttpClient - Fix #551
Improve components serialization to better support workflow interruption - PR #543 by @tryvin
Improve components serialization to better support workflow interruption - PR #543 by @tryvin
Merged: - PR #546 by @plusemon - PR #540 by @viicslen - PR #534 by @saithink
Merged:
Fix #544 - get schema in SQL Tools
Fix #544 - get schema in SQL Tools
Support Anthropic image as a file ID
Fix Mistral token usage reporting
Fix Mistral token usage reporting
Nothing published for this version
Support OpenSearch 3.6.0 or higher
Fix error handler in case MissingCallbackParameter
MissingCallbackParameterIn the process of fixing a bug ( #523 ) we worked on a more efficient algorithm to manage the cut of the message history if it goes out of the context
In the process of fixing a bug ( #523 ) we worked on a more efficient algorithm to manage the cut of the message history if it goes out of the context window. The intent was to minimize the context loss when cutting.
The context window cut must be considered the last resort to prevent fatal errors with provider APIs. Before reaching the limit of the context window you should use summarization to compact the context into a single message as a new starting point.
In order to handle all edge cases, the internal trimmer can identify a cutting point slightly less aggressive than the initially identified. To make sure the agent conversation stays in the limit, you should configure the context window limit in the agent chat history with a margin of 5%-10% from the actual limit of the underlying model.
If your model works with a 200K context window, you should instantiate your chat history with 190K for example.
We introduced a couple of built-in rules for Ai as a judge assertions:
https://docs.neuron-ai.dev/agent/evaluation#ai-as-a-judge
Nothing published for this version
`php class MyAgent extends Agent { protected function provider(): AIProviderInterface { return new Anthropic( key: 'ANTHROPIC_KEY', model: 'ANTHROPIC_
PR #482 by @hudhaifas
class MyAgent extends Agent
{
protected function provider(): AIProviderInterface
{
return new Anthropic(
key: 'ANTHROPIC_KEY',
model: 'ANTHROPIC_MODEL'
)->systemPromptBlocks([
['type' => 'text', 'text' => 'Static instructions...', 'cache_control' => ['type' => 'ephemeral']],
['type' => 'text', 'text' => 'Dynamic context...']
]);
}
}
Fix toolkit guidelines in instructions - PR #521 by @tryvin
Fix toolkit guidelines in instructions - PR #521 by @tryvin
Fix message mapper for single block - PR #518 by @qwertyquest
Fix message mapper for single block - PR #518 by @qwertyquest
Track stop reason when streaming in Gemini - PR #516 by @joecharm
Track stop reason when streaming in Gemini - PR #516 by @joecharm
fix monitoring message reporting
fix monitoring message reporting
Nothing published for this version
Evaluation skills to help AI coding assistant implement high quality evaluators.
Evaluation skills to help AI coding assistant implement high quality evaluators.
The evaluation module uses a PHP configuration file to control how evaluation results are output. The config system supports multiple output drivers,
The evaluation module uses a PHP configuration file to control how evaluation results are output. The config system supports multiple output drivers, enabling results to be sent to console, files, databases, or external APIs simultaneously.
Check out the official documentation.
Lower minimum guzzle version. This allow broader compatibility.
Lower minimum guzzle version. This allow broader compatibility.
You can now delete documents only by sourceType instead of being forced to use sourceName too - PR #503 by @bytestream
deleteBy for vector storeYou can now delete documents only by sourceType instead of being forced to use sourceName too - PR #503 by @bytestream
Distribute Agent Skills: https://docs.neuron-ai.dev/overview/agentic-development
Distribute Agent Skills: https://docs.neuron-ai.dev/overview/agentic-development
One column per month.