import { morph } from 'morph'Better, faster, cheaper search via subagents
Use cases
Code Search · Log Search · Package Search. Read the docs →
const result = await morph.warpGrep.execute({ query: 'Find where billing invoices are emailed', repoRoot: './my-project',}) // Clean results, no context pollutionconsole.log(result.contexts) // Relevant code onlyconsole.log(result.summary) // Search findingsSearch codebases without
blowing context
Isolated context window
WarpGrep is a specialized AI model that searches your codebase in a separate context window. Instead of your main agent running grep and accumulating context, WarpGrep does intelligent multi-step searching and returns clean results.
import { morph } from 'morph' // WarpGrep searches in isolationconst result = await morph.warpGrep.execute({ query: 'Find where billing invoices are generated', repoRoot: '.', excludes: ['node_modules', '.git'], includes: ['src/**/*.ts']}) // Returns only relevant code sectionsfor (const ctx of result.contexts) { console.log(ctx.file, ctx.content)}Multi-turn reasoning
Your agent calls WarpGrep with a natural language query. WarpGrep reasons about what to search, runs up to 12 parallel tool calls (grep, read, list_dir) across 3 turns in under 6 seconds, and returns the relevant code sections. No embeddings required.
Best-in-class F1 score
WarpGrep achieves 0.73 F1 in just 3.8 steps—3x fewer than comparable agentic approaches. Deep logic queries show 2x-6x better performance.
WarpGrep achieves 0.73 F1 in 3.8 steps — 3x fewer than comparable agentic approaches
How WarpGrep works
Parallel breadth-first search
