Quick Start

Rohrpost compresses long-running LLM conversation context into compact capsules — structured briefs that preserve essential information while drastically reducing token usage.

Prerequisites

  • Rust toolchain (1.75+) or a pre-built binary from the releases page
  • An LLM provider API key (OpenAI, Anthropic, or any OpenAI-compatible endpoint)

Install the CLI

cargo install rohrpost-cli

Or download a pre-built binary for your platform from the latest release.

Create your first capsule

# Initialize a new Rohrpost store
rohrpost init my-project

# Ingest a conversation transcript
rohrpost ingest --file conversation.json --project my-project

Rohrpost will extract key decisions, code references, and architectural context from the conversation and compress it into a capsule.

Query compressed context

# Retrieve a brief for your next LLM call
rohrpost recall --project my-project --query "authentication flow"

The result is a token-efficient brief you can inject directly into your next prompt, typically achieving 60–80% token reduction compared to replaying the full conversation history.

Next steps

  • Learn about Capsule Architecture to understand how compression works
  • Configure distillation settings for your specific use case