ScaleDown offers langchain-scaledown
Product
LangChain middleware that adds ScaleDown small language models to agents for summarization, compression and extraction.
- Currently Offered
LangChain middleware that adds ScaleDown small language models to agents for summarization, compression and extraction.
langchain-scaledown adds ScaleDown small language models to any LangChain agent as middleware. Agents use it to summarize long chats, compress retrieved documents or turn conversations into structured records instead of spending frontier-model tokens.
The package runs inside the developer's application, has no telemetry and sends only the text each middleware needs to ScaleDown's API endpoints. Its summarization middleware is a drop-in for LangChain's built-in SummarizationMiddleware.
A unified OpenAI-compatible API and gateway for accessing hundreds of LLMs from many providers.
OpenAI-compatible API proxy with built-in conversational memory supporting 300+ models.
Async inference API for long-horizon AI agents, serving leading open-source models through OpenAI- and Anthropic-compatible endpoints.
Developer API for Anthropic's Claude family of frontier language models including Opus, Sonnet, and Haiku.
Schema-based structured data extraction from documents with confidence scores.
Pre-optimized APIs for Kimi, DeepSeek, GLM on Baseten Inference Stack.
A unified AI gateway and LLM router exposing one OpenAI-compatible endpoint across 600+ models.
API that trains compact reusable memories on organizational context for agent inference.
Developer API providing pay-per-token access to DeepSeek's hosted language models.
Conifer Gateway is one OpenAI- and Anthropic-compatible endpoint for more than 200 models.
ZeroEntropy's zembed-1 embedding model converts documents into dense vectors for semantic search and retrieval-augmented generation.
Poe is a fast AI chat platform providing access to multiple top language models.
A unified OpenAI-compatible API and gateway for accessing hundreds of LLMs from many providers.
OpenAI-compatible API proxy with built-in conversational memory supporting 300+ models.
Adaptive OpenAI-compatible endpoint coordinating open-weight models for chat and agents.
Async inference API for long-horizon AI agents, serving leading open-source models through OpenAI- and Anthropic-compatible endpoints.
Developer API for Anthropic's Claude family of frontier language models including Opus, Sonnet, and Haiku.
Schema-based structured data extraction from documents with confidence scores.
Pre-optimized APIs for Kimi, DeepSeek, GLM on Baseten Inference Stack.
Production LLM applications built with LangChain and LangGraph.
A unified AI gateway and LLM router exposing one OpenAI-compatible endpoint across 600+ models.
API that trains compact reusable memories on organizational context for agent inference.
Developer API providing pay-per-token access to DeepSeek's hosted language models.
Conifer Gateway is one OpenAI- and Anthropic-compatible endpoint for more than 200 models.
ZeroEntropy's zembed-1 embedding model converts documents into dense vectors for semantic search and retrieval-augmented generation.
Poe is a fast AI chat platform providing access to multiple top language models.
Retrieval layer grounding agent answers in your own documents.