Download the PHP package cable8mm/nano-ai without Composer
On this page you can find all versions of the php package cable8mm/nano-ai. It is possible to download/install these versions without Composer. Possible dependencies are resolved automatically.
Download cable8mm/nano-ai
More information about cable8mm/nano-ai
Files in cable8mm/nano-ai
Package nano-ai
Short Description A lightweight, zero-configuration PHP SDK for straightforward text generation and multimodal AI interactions.
License MIT
Homepage https://github.com/cable8mm/nano-ai
FAQ
Example:
In general, it is recommended to use always a project to download your libraries. In an application normally there is more than one library needed.
- Some hosting areas are not accessible by a terminal or SSH. Then it is not possible to use Composer.
- To use Composer is sometimes complicated. Especially for beginners.
- Composer needs much resources. Sometimes they are not available on a simple webspace.
- If you are using private repositories you don't need to share your credentials. You can set up everything on our site and then you provide a simple download link to your team member.
- Simplify your Composer build process. Use our own command line tool to download the vendor folder as binary. This makes your build process faster and you don't need to expose your credentials for private repositories.
Informations about the package nano-ai
NanoAI
Ultra-lightweight, zero-config PHP AI SDK. No agents/RAG — just generate() for text + image (multimodal) calls.
Use it for testing, lightweight pipelines, and idea validation.
Supported Providers
| Provider | Description |
|---|---|
openai |
Direct connection to OpenAI Chat Completions |
openrouter |
Call most models (OpenAI/Gemini/DeepSeek/Qwen, etc.) via a single API by just changing the model name. Free models use :free suffix |
Installation
Usage
Configuration Options
The Client constructor accepts an optional $options array to customize behavior:
Available Options
| Option | Type | Default | Description |
|---|---|---|---|
timeout |
int |
120 |
Maximum time to wait for API response (seconds) |
connectTimeout |
int |
30 |
Maximum time to establish connection (seconds) |
baseUrl |
string |
Provider default | Override the API base URL |
referer |
string |
None | Custom Referer header sent with requests |
title |
string |
None | Custom X-Title header sent with requests |
Note: Options like baseUrl, referer, and title are particularly useful when using OpenRouter or custom OpenAI-compatible APIs.
To test with free/low-cost models:
Exceptions
AuthenticationException— Missing/invalid API Key (401/403)RateLimitException— Request rate limit exceeded (429), common with free modelsApiException— Other API errors, providesgetStatusCode()/getResponseBody()NetworkException— cURL-level failures such as timeouts
All extend NanoAIException, so you can catch them all at once.
Using Guzzle as the HTTP Client (Optional)
By default, nano-ai uses a minimal cURL-based HTTP client with zero external dependencies.
If you prefer Guzzle (or need its advanced features like middleware, retries, proxies, etc.),
install it and inject the provided GuzzleHttpClient:
The GuzzleHttpClient implements the same HttpClientInterface as the default cURL client,
so it is a drop-in replacement. All nano-ai features (timeouts, error handling, etc.) work identically.
Adding a New Provider
Create a class implementing NanoAI\Provider\ProviderInterface (if it's an OpenAI-compatible API, you can extend AbstractOpenAICompatibleProvider), then add one line to ProviderFactory::make().
No need to modify Client or HttpClient.
Development
End-to-End Testing (Real API)
The default test suite uses mock servers and fake HTTP clients — no external network access required.
To verify against the real OpenAI/OpenRouter APIs:
These tests are opt-in and skipped by default. They make real API calls
and may incur costs. The tests/config.json file is gitignored and will
never be committed.
License
MIT
All versions of nano-ai with dependencies
ext-curl Version *
ext-json Version *