Functions

A filter that truncates chat history to retain the latest messages while preserving the system prompt for optimal context management.

Takes a promt an makes a sctructured plan and folows it orderly, uses consolidation and ranking techniques to save on context and provided the mos relevant per step results

Manage context use by summarizing conversation history.

Determines what information to store in memory and what memories to add to the context, using openai to determine relevance

Truncate chat context length with 'token limit' and 'max turns', showing status while limit exceeded.

A comprehensive context window tracker and metrics dashboard for OpenWebUI that provides real-time feedback on token usage, cost tracking, and performance metrics for all major LLM models.

This pipe significantly augments the intelligence of an LLM with CoT (Chain of Thought) reasoning and context window management. It first generates an outline for the input, and then answers each part of the outline separately.

a filter to cut the context of historical conversation turns (Support for head and tail cutting)

LLMs usually demonstrate poor performance when the context window is largely occupied. This pipe is designed to handle long context by splitting the context into multiple chunks, and then concatenating the results.

Adds the whole document to context

Truncates chat context length with token limit and max turns, logging the chat turn data.

Gives model current date and time context for each message. Don't forget to adjust the timezone in the settings.

A filter to remove markdown images in chat so they don't interfere with the context. Useful for people using their own tools and functions to generate images.

A filter that truncates chat history to retain the 7000 tokens while preserving the system prompt for optimal context management.

add user info to context without having to use a tool

A filter that truncates chat history to retain the latest n-th user and assistant messages while always keeping the system prompt and also first message pair.

Reduces context to X characters - and automatically deletes the old stuff in conversation.

Local mutil-LLM context-enhancing pre-processor.

Brave based web-search with query optimizer based on last chat messages

Enhances Ollama API facade interactions by providing additional metadata including id, session_id, and chat_id for seamless integration with applications requiring rich context information.

A Pipe that enriches the pipelines server by including additional metadata fields such as id, chat_id, and session_id for advanced context handling. You should disable the normal pipelines connection if you use this.

Log your inputs and completions to Weave (by Weights & Biases) for LLMOps / observability

Removes all context from the request to the model, leaving only the last user message (useful for cases where previous messages are not needed).

Filters messages to keep the system prompt and the last N non-system messages, optionally further truncating based on a maximum character count limit.

A filter for redacting various personally identifiable information (PII) from a nordic context

Takes a promt an makes a sctructured plan and folows it orderly, uses consolidation and ranking techniques to save on context and provided the mos relevant per step results