Documentation
¶
Overview ¶
Package providers holds LLM provider adapters. The ModelClient contract itself lives in package pipeline (to avoid an import cycle) and is aliased here so `providers.ModelClient` is the name callers use.
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type ModelClient ¶
type ModelClient = pipeline.ModelClient
ModelClient abstracts a single LLM provider/model endpoint. Implementations must NOT be assumed to be Anthropic-specific anywhere outside this package and its subdirectories.
Source Files
Directories
¶
| Path | Synopsis |
|---|---|
|
Package anthropic implements pipeline.ModelClient against the Anthropic Messages API using nothing but net/http and encoding/json.
|
Package anthropic implements pipeline.ModelClient against the Anthropic Messages API using nothing but net/http and encoding/json. |
|
Package cli adapts any command-line coding agent into a pipeline.ModelClient by running it as a subprocess.
|
Package cli adapts any command-line coding agent into a pipeline.ModelClient by running it as a subprocess. |
|
Package mock provides a ModelClient test double.
|
Package mock provides a ModelClient test double. |
|
Package openai implements pipeline.ModelClient against the OpenAI chat-completions API — and, because that shape has become a de facto standard, against everything that speaks it: Ollama, vLLM, llama.cpp, Groq, Together, OpenRouter, LM Studio and Azure OpenAI.
|
Package openai implements pipeline.ModelClient against the OpenAI chat-completions API — and, because that shape has become a de facto standard, against everything that speaks it: Ollama, vLLM, llama.cpp, Groq, Together, OpenRouter, LM Studio and Azure OpenAI. |
Click to show internal directories.
Click to hide internal directories.