Vexp — Context Engine for AI Coding Agents
Context Engine · v2.0.25

Build a local dependency graph. Serve precise context to every AI agent.

Vexp indexes your codebase into a graph of files, functions, and imports — then feeds only the relevant slice to your AI coding agent. Same setup in VS Code and the CLI.

Runs 100% locally Works in VS Code & CLI Multi-repo ready
vexp · cli
65–70% token savings
nicola@localhost:~/projects/augenscapital$ vexp
Context Engine for AI Coding Agents
Daemon running
LLM vexp-devmind-v1
Nodes 208
Edges 81
Files 126
Uptime 4m
1 Setup & Config
2 Explore & Analyze
3 Token Savings
4 License
vexp>
208
Nodes Indexed
111.2K
Tokens Saved / 30d
84.2%
Avg. Token Savings
4m
Time To First Index
Key Features

Everything your agent needs. Nothing it doesn't.

Vexp keeps a live graph of your repo, so context is precise instead of a guess.

01

Dependency Graph Indexing

Scans files, functions, and imports into nodes and edges — refreshed as your code changes.

02

Multi-Repo Workspaces

Link several repos with one workspace.json. Pull context from projects you haven't opened.

03

Token Savings Tracking

Every context call is logged. See saved tokens, average %, and export a full report.

04

Local LLM Compression

Optional vexp-devmind model runs on your machine. Zero data ever leaves it.

05

VS Code + CLI Parity

Every panel action has a matching CLI command — switch tools without switching workflows.

06

Context Capsules

Generate a packaged snapshot of graph context to hand off to an agent or teammate.

Multi-Repo

Context from every repo — even the ones you haven't opened

Link multiple repositories into one Vexp workspace, so your agent pulls context across projects that aren't open in the editor right now.

dev
3.0k nodes · 185 files
dev-web
370 nodes · 160 files
dev-test
14 nodes · 2 files
{
  "name": "dev",
  "repos": [
    { "alias": "dev-web", "path": "../dev-web" },
    { "alias": "dev-test", "path": "../dev-test" }
  ]
}
Token Savings

Savings you can see, everywhere you work

Every run_pipeline call's savings, tracked locally and surfaced in your editor or terminal.

111.2K
Total Saved (30d)
84.2%
Avg Saving
13
Total Calls
3.5GB
Local LLM Size

Give your AI agent the context it's been missing

Index your first repo in minutes. Works in VS Code and your terminal, right now.

Start Indexing Free
Vexp · Context Engine for AI Coding Agents

Post a Comment

0 Comments