neograph::llm::OpenAIProvider Class Reference

LLM provider for OpenAI-compatible APIs. More...

#include <openai_provider.h>

Inheritance diagram for neograph::llm::OpenAIProvider:

Collaboration diagram for neograph::llm::OpenAIProvider:

Classes
struct	Config
	Configuration for OpenAI-compatible API connections. More...

Public Member Functions
asio::awaitable< ChatCompletion >	complete_async (const CompletionParams &params) override
	Async completion — dispatches over the owned ConnPool so successive calls reuse a kept-alive TCP+TLS connection.
ChatCompletion	complete_stream (const CompletionParams &params, const StreamCallback &on_chunk) override
	Sync completion is inherited from Provider::complete(), which drives complete_async via neograph::async::run_sync.
std::string	get_name () const override
	Get the provider name (e.g., "openai", "claude").
asio::awaitable< ChatCompletion >	invoke (const CompletionParams &params, StreamCallback on_chunk) override
	v1.0 single-dispatch override (Candidate 6 PR6).
	~OpenAIProvider ()
	Destructor — shuts down the long-lived HTTP loop + worker thread held alongside the ConnPool.
Public Member Functions inherited from neograph::Provider
virtual ChatCompletion	complete (const CompletionParams &params)
	Perform a synchronous LLM completion.
virtual asio::awaitable< ChatCompletion >	complete_stream_async (const CompletionParams &params, const StreamCallback &on_chunk)
	Async streaming completion.

Static Public Member Functions
static std::unique_ptr< OpenAIProvider >	create (const Config &config)
	Create an OpenAI provider instance.
static std::shared_ptr< Provider >	create_shared (const Config &config)
	Same as create but returns a shared_ptr<Provider>.

Detailed Description

LLM provider for OpenAI-compatible APIs.

Connects to any endpoint following the OpenAI chat completions API format (/v1/chat/completions). Configure via Config struct.

auto provider = OpenAIProvider::create({
    .api_key = "sk-...",
    .default_model = "gpt-4o-mini"
});

See also

SchemaProvider for multi-vendor support (Claude, Gemini).

Definition at line 38 of file openai_provider.h.

Member Function Documentation

complete_async()

asio::awaitable< ChatCompletion > neograph::llm::OpenAIProvider::complete_async ( const CompletionParams &  params )

overridevirtual

Async completion — dispatches over the owned ConnPool so successive calls reuse a kept-alive TCP+TLS connection.

The previous "fresh socket per call" comment has aged out: the pool is bound to a long-lived background io_context owned by this provider, so run_sync per-call destruction no longer affects pool lifetime.

Reimplemented from neograph::Provider.

create()

static std::unique_ptr< OpenAIProvider > neograph::llm::OpenAIProvider::create ( const Config &  config )

static

Create an OpenAI provider instance.

Parameters

config

Connection and authentication configuration.

Returns

A unique_ptr to the created provider.

create_shared()

static std::shared_ptr< Provider > neograph::llm::OpenAIProvider::create_shared ( const Config &  config )

static

Same as create but returns a shared_ptr<Provider>.

Useful when one provider needs to be reused across multiple NodeFactory closures, A2A servers, or threads — capturing a unique_ptr into a copyable lambda is impossible without an explicit release(). This peer drops the boilerplate.

get_name()

std::string neograph::llm::OpenAIProvider::get_name ( ) const

inlineoverridevirtual

Get the provider name (e.g., "openai", "claude").

Opaque debug identifier, not a typed dispatch key. Different subclasses pick different conventions:

• OpenAIProvider always returns "openai".
• SchemaProvider returns whatever's in the schema's name field — could be "openai", "claude", "openai-responses", "gemini", or a user-defined schema id.
• RateLimitedProvider delegates to its inner provider.

Code branching on the exact string (e.g. if (get_name() == "openai-responses")) is brittle — a custom schema named "openai-responses-v2" slips through, or a future rename silently breaks the branch. Use it for logging, telemetry, or version-pinning diagnostics. For typed behaviour, add a typed ProviderKind accessor or branch on the schema's actual fields.

Returns

Opaque provider identifier string.

Implements neograph::Provider.

Definition at line 104 of file openai_provider.h.

invoke()

asio::awaitable< ChatCompletion > neograph::llm::OpenAIProvider::invoke ( const CompletionParams &  params, StreamCallback  on_chunk  )

overridevirtual

v1.0 single-dispatch override (Candidate 6 PR6).

Native invoke() that directly drives the ConnPool — no extra hop through complete_stream_async's default worker-thread bridge for streaming, no extra hop through complete_async for non- streaming. The 4 legacy overrides above stay as thin adapters over invoke() through the v0.9 deprecation window; v1.0 deletes them along with the base-class virtuals.

Reimplemented from neograph::Provider.

---

The documentation for this class was generated from the following file:

• include/neograph/llm/openai_provider.h