Roka Roka Docs

Documentation

Welcome to Roka

Roka is a context-pruning layer for AI coding agents. It collapses noisy logs, force-keeps real errors, and packs the result into a token budget — before the model ever sees the dump.

Local prune tools do not upload your logs. MCP connect / serve / watch only call the API to verify a Pro key.

What Roka is

Point the CLI at a file or pipe stdin. Or connect an MCP-compatible agent so it can call prune_file, prune_logs, and prune_tail itself. Output is always the pruned context, not a summary written by a model.

cat deploy.log | roka prune --query "what broke" --budget 8000

Get started

Install the CLI

One-line install for macOS, Linux, and Windows (WSL2). No sudo. Checksum-verified binary.

Connect an agent

Pro: register roka-mcp with Claude Code, Cursor, Codex, Copilot, Windsurf, VS Code, or Cline.

CLI usage

Flags, stdin/file/output, query + budget, and the Free vs Pro path.

MCP tools

Exact tool names, arguments, defaults, and what the agent gets back.

How pruning works

Collapse → rank / force-keep → fit. Crash regex, default budget, what stays local.

FAQ

Offline use, quotas, Claude uploads vs Roka, RAG, data retention.

Quick links

Install CLI →