MorphMorph
Morph WarpGrep

Better, faster, cheaper search via subagents

A code search subagent that runs in a separate context window. Up to 36 tool calls across 3 turns—no embeddings required.
Try it nowSee benchmarksSee Docs

Use cases

Code Search · Log Search · Package Search. Read the docs →

01
import { morph } from 'morph'
02
03
const result = await morph.warpGrep.execute({
04
  query: 'Find where billing invoices are emailed',
05
  repoRoot: './my-project',
06
})
07
08
// Clean results, no context pollution
09
console.log(result.contexts)  // Relevant code only
10
console.log(result.summary)   // Search findings
WarpGrep

Search codebases without blowing context

Isolated context window

Searches your codebase in a separate context window. Your main agent only sees clean results—never raw grep output or dead-end searches.

01
import { morph } from 'morph'
02
03
// WarpGrep searches in isolation
04
const result = await morph.warpGrep.execute({
05
  query: 'Find where billing invoices are generated',
06
  repoRoot: '.',
07
  excludes: ['node_modules', '.git'],
08
  includes: ['src/**/*.ts']
09
})
10
11
// Returns only relevant code sections
12
for (const ctx of result.contexts) {
13
  console.log(ctx.file, ctx.content)
14
}

Multi-turn reasoning

Up to 36 parallel tool calls across 3 turns. Reasons about what to search, runs grep/read/list_dir in parallel, returns relevant code in seconds.

WarpGrep
Turn 1 · 8 parallel calls
├─grep -r "webhook.*handler"2.1k
│ ⎿
├─grep -E "async.*subscribe"1.8k
│ ⎿
├─read src/api/billing.ts:1423.2k
│ ⎿
├─grep "customer\.create"1.4k
│ ⎿
├─list_dir src/services/0.8k
│ ⎿
├─read src/lib/stripe.ts:1-802.6k
│ ⎿
├─grep -r "invoiceId"1.9k
│ ⎿
└─read src/api/checkout.ts:552.3k
⎿
14 tool calls · 3 turns~3.4s

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.

1Warp Grep
0.733.8 steps
2SWE Grep
0.723.7 steps
3Claude Haiku
0.7212.4 steps
4Gemini Flash
0.6610.8 steps
5GLM 4.5
0.3414.5 steps
6mgrep
0.181 steps

WarpGrep achieves 0.73 F1 in 3.8 steps — 3x fewer than comparable agentic approaches

How WarpGrep works

1/9
1User Query
Input
"How does auth work in this repo?"

Parallel breadth-first search

Parallel breadth-first search
GitHub Search

Search any GitHub repo. No clone required.

Point WarpGrep at any public GitHub repository and search it remotely. Ask complex questions about unfamiliar codebases without downloading a single file.

Search pytorch/pytorch, facebook/react, or any repo by name

Pin specific branches or tags for version-locked search

Works as an MCP tool, SDK call, or in the playground

Try it in the playground
github_codebase_search
WarpGrep searching pytorch/pytorch on GitHub without cloning

Ship your first search workflow in minutes.

Try it now

$10/month in free compute