Search code and notes.
Build context for your task.
Give your agent the code behind the task.
Not a longer result list. A bounded pack connecting current work to code an agent needs to understand before editing.
Current work becomes retrieval signal
ig context --since main finds the Git merge base of the current branch and main. It includes committed changes from that base to the current branch, plus staged, unstaged, and untracked files, task anchors, and paths found in an issue or stack trace. The main ref is not itself the diff base. Context packs follow dependencies, dependents, definitions, callers, references, tests, configuration, docs, and recent co-changes.
Each item includes exact lines, its role, why it was selected, and how it relates to the task.
$ ig context "fix refresh-token races" --since main --budget 8000
# ivygrep context
...
Budget: 7642 / 8000 estimated tokens
Coverage: 7 files | 2 primary | 1 definitions | 1 dependencies | 0 dependents | 2 callers | 0 references | 1 tests | 0 config | 0 docs
Candidates: 31 retrieved | 14 selected
## Evidence
### 1. src/auth/refresh.rs:118-166 [primary, definition]
Why: task anchor; changed implementation.
Signals: lexical, symbol, git change.
Use current changes
Context packs include staged, unstaged, and untracked changes. With --since main, they also include commits on the current branch since the merge base of that branch and main.
Include related code
Each pack can include definitions, callers, dependencies, tests, configuration, and documentation. Selection reasons explain each item.
Set a token budget
The --budget option limits the rendered pack in estimated tokens. Context packs supply evidence for an agent to inspect.
Search notes by meaning
Markdown vaults, research logs, meeting notes, decisions, and agent memories stay local.
Local semantic recall
Index your notes once. Search them from the CLI, MCP, Web UI, or TUI. A filesystem watcher updates changed files.
74.9% recall@20 at 87.63 ms warm p95 on MemoryQuest (v1.2.7).
This historical benchmark measures retrieval. It does not measure answer accuracy or the current model.
# Build local semantic index once. $ ig --add ~/notes --wait-for-enhancement # Reuse it for memory retrieval. $ ig -n 20 \ "what did we decide about cache invalidation?" \ ~/notes
Local vaults
Search Markdown, text, JSON, and other supported documents without uploading notes or queries.
Updated notes
The filesystem watcher indexes changed files. Background workers update their hash vectors and neural vectors.
Inspect evidence
Results preserve exact file paths, lines, scores, and previews so retrieved memories stay auditable.
Built for live knowledge, not one-off searches
Task state, repository relationships, notes, and incremental indexes stay connected across CLI, MCP, and Web.
Task-aware inputs
Combine a task, issue, trace, branch diff, staged files, and dirty worktree before selecting code.
Repository relationships
Expand from task anchors through definitions, callers, references, dependencies, dependents, tests, configuration, docs, and recent co-changes.
Bounded evidence
Token-aware selection and deduplication produce one pack with exact paths, lines, roles, reasons, and relationships.
Thin worktree overlays
Worktrees reuse the repository base index. Their overlays store changed chunks and tombstones that hide base files deleted on that branch.
Incremental availability
Changed chunks update incrementally. Starting with v1.3.1, files confirmed missing disappear from results immediately while the watcher removes owned index entries or records an overlay tombstone. Permission and other read errors are not treated as deletion.
Exact and semantic
Move between semantic notes, natural-language code intent, literal text, symbols, references, callers, paths, and project-wide search.
One context contract
CLI, JSON, MCP, and Web expose the same structured search results and task-context packs.
Local by design
No hosted inference or telemetry. Code, notes, queries, embeddings, and indexes stay on your machine. Neural search downloads pinned model assets on first use; --hash avoids the download.
Browse indexed workspaces
The local browser UI searches tracked workspaces, scopes results to a folder, and opens indexed files beside the result list.
Result cards show language icons, syntax-highlighted snippets, line markers, and selected-state highlighting. The file viewer highlights focused result lines and previews Markdown with a source toggle.
Every bind, loopback included, prints an authenticated URL, so other users on a shared machine cannot query the local API. Open that URL directly and do not share it: the browser exchanges the token for an HttpOnly session cookie and removes it from the address, while a bare http://127.0.0.1:4747/ returns 401. Set IVYGREP_NO_BROWSER=1 to skip the automatic browser launch. On another device, replace only 127.0.0.1 with the server's LAN or Tailscale address and preserve the complete token=... query string. Restart the daemon to invalidate access.
Non-loopback serving uses plain HTTP: the token authenticates requests but does not encrypt source or results. Use a trusted network only, preferably Tailscale or an SSH tunnel. Never expose the port directly to the internet or untrusted Wi-Fi.
$ ig --web $ ig --web "auth flow" . $ ig --web --host 0.0.0.0 --port 4747
How indexing and search fit together
BM25 and literal search are available first; background jobs add hash and neural vectors.
Lexical index
File chunks go into SQLite and Tantivy first, so search works before vector jobs finish.
Rank matching files
Search combines exact text, keywords, symbols, paths, and vector matches. Results include file paths, line numbers, and previews.
Local vectors
Background workers add hash vectors and neural vectors. The default potion-code-16m-v2 model downloads 33.5 MB of assets and runs on CPU. Metal and CUDA accelerate optional transformer profiles.
Git-aware indexing
Worktrees share a base index. Branch switches reconcile only changed paths, including deletions.
Worktree overlays
Each worktree references the shared base index. Its overlay stores changed chunks and tombstones that hide base files deleted on that branch.
Branch-switch deltas
Merkle reconciliation handles changed paths, including deletions. Content hashes skip byte-identical files across branches.
Context stays practical at repository scale
Three hash-only trials on deterministic synthetic CC0 data measure local indexing, query latency, and footprint at scale.
6.19 ms
Warm CLI p95
150,576 chunks/s
Controlled indexing
0.42 GiB
Final index size
Median of three sequential v1.2.7 trials on deterministic synthetic CC0 data in hash-only mode. These are scale and footprint measurements, not semantic quality or agent outcomes. Hardware, repository shape, index state, and load affect absolute results.
Current model cost and concurrent latency Historical scale results (v1.2.7) Public million-chunk benchmark Benchmark dashboard
45 language/file types, 24 Tree-sitter ASTs
Tree-sitter sources get function/class-level chunks. Other supported files use heuristic chunks.
Tree-sitter AST 24 languages
Heuristic chunking 21 more
MCP for coding agents
Use MCP search to find code. Set output=context_pack, optional since=main, and a token budget to build the same context pack available from CLI and Web.
One-command setup
The installer preserves existing configuration, verifies MCP, and runs a search.
Claude Code
ig agent install claude
Codex
ig agent install codex
Cursor
ig agent install cursor
ig agent doctor
Manual MCP setup
Claude Code
claude mcp add -s user ig -- ig --mcp
Codex
codex mcp add ig -- ig --mcp
Gemini CLI
gemini mcp add --scope user --transport stdio ig ig --mcp
Cursor
{"mcpServers":{"ig":{"type": "stdio", "command": "ig", "args": ["--mcp"]}}}
OpenCode
{"mcp":{"ig":{"type": "local", "command": ["ig", "--mcp"], "enabled": true}}}
MCP search settings
Pass the absolute repository or worktree path to ig_search. Use a natural-language query for concepts, literal=true for identifiers, or output=context_pack with since=main and budget_tokens=8000 for implementation context.
The first search creates an index and starts incremental watching. Worktrees reuse one shared base index and store only divergent chunks and tombstones.
MCP search flow
Contributing
Choose a scoped issue, run the relevant checks, and keep the pull request focused.
Discuss larger changes
Use Discussions for questions and designs that affect storage, ranking, or compatibility.
DiscussionsOpen a pull request
Follow the contributor guide and include the checks that cover the change.
Contributor guideInstall ivygrep
The shell installer selects CUDA only on compatible NVIDIA Linux x86_64 hosts with the CUDA 13 runtime, and Metal on Apple Silicon. Homebrew uses portable builds on Linux and Intel macOS, and Metal on Apple Silicon. GPU builds only speed up transformer profiles (IVYGREP_MODEL_PROFILE=code|code-hq|general); the default potion-code-16m-v2 profile runs on CPU. Run ig hardware to inspect the selected build.
Homebrew for macOS and Linux
$ brew install bvolpato/tap/ivygrep
Metal on Apple Silicon for transformer profiles. Portable on Linux and Intel macOS.
Shell installer for Linux and macOS
$ curl -fsSL https://raw.githubusercontent.com/bvolpato/ivygrep/main/install.sh | sh
Selects CUDA only on Linux x86_64 with a supported NVIDIA GPU and complete CUDA 13 runtime; selects Metal on Apple Silicon. Other supported hosts use the portable build.
WinGet for Windows
$ winget install --id BrunoVolpato.ivygrep --exact
Portable Windows build. Upgrade with winget upgrade --id BrunoVolpato.ivygrep --exact.
Windows PowerShell
irm https://raw.githubusercontent.com/bvolpato/ivygrep/main/install.ps1 | iex
Portable Windows build. Verifies checksum and updates PATH.
Search once. Build context once.
Find code by intent
ig "where is refresh token rotated?"
Build bounded task context
ig context "fix refresh-token races" --since main --budget 8000