omni

From public sources

Your agent pays twice for output it has already seen. OMNI returns a handle instead: 97.2% off a file read twice. Nothing deleted, nothing invented.

CLI toolCategory: Developer Tools · AI codingOpen sourcedata checked Oct 16, 2026

Add to compare Add to collection

Overview

agentic-ai
ai-agents
antigravity
claude-code
cli
context-compression
context-distillation
context-engineering
cost-reduction
cursor
developer-tools
homebrew

Installation

Commands are shown, never run by the marketplace. Check them against the official documentation.

Official websiteFrom public sourcesfrom catalog metadata
omni.weekndlabs.com/docs ↗
Official release filesFrom public sourcesfrom the latest official release
github.com/fajarhide/contexere/releases/tag/v0.8.1 ↗
Language package registryFrom public sourcesfrom the package manifest — publication on the registry is not verified
cargo install omni
Not verified: check that this package in the registry belongs to this project before installing.
crates.io/crates/omni ↗
Build from sourceFrom public sourcesfrom catalog metadata
Build from source: no validated build recipe — follow the repository instructions.
github.com/fajarhide/omni ↗

Compatibility

Unknown means no evidence was found — not that the product is incompatible.

Platform	Status	Architecture	Evidence	Confidence
Windows	Compatible	x86_64	official build in release v0.8.1: omni-v0.8.1-x86_64-pc-windows-msvc.zip	0.70
macOS	Compatible	arm64	official build in release v0.8.1: omni-v0.8.1-aarch64-apple-darwin.tar.gz	0.70
	Compatible	x86_64	official build in release v0.8.1: omni-v0.8.1-x86_64-apple-darwin.tar.gz	0.70
Linux	Compatible	arm64	official build in release v0.8.1: omni-v0.8.1-aarch64-unknown-linux-musl.tar.gz	0.70
	Compatible	x86_64	official build in release v0.8.1: omni-v0.8.1-x86_64-unknown-linux-musl.tar.gz	0.70
Container (OCI)	Unknown	—	—	—

Releases

checked Oct 30, 2026

v0.8.1LatestOct 29, 2026Official source ↗

Files in 0.8.1

The marketplace does not host these files: downloads go to the official source. Files have not been scanned by us.

File	Platform	Size	SHA-256	Download from the official source
omni-v0.8.1-aarch64-apple-darwin.tar.gz	macos · arm64 · tarball	4.1 MB	c40b4cf6e831…	Official source ↗
omni-v0.8.1-aarch64-unknown-linux-musl.tar.gz	linux · arm64 · tarball	4.3 MB	8e2419bbe2bb…	Official source ↗
omni-v0.8.1-x86_64-apple-darwin.tar.gz	macos · x86_64 · tarball	4.4 MB	50f014d3563c…	Official source ↗
omni-v0.8.1-x86_64-pc-windows-msvc.zip	windows · x86_64 · zip	8.8 MB	72fc76e87169…	Official source ↗
omni-v0.8.1-x86_64-unknown-linux-musl.tar.gz	linux · x86_64 · tarball	4.7 MB	4e902f568d56…	Official source ↗
SHA256SUMS	checksums	539 B	d13b919c5d99…	Official source ↗

Release history

v0.8.0Oct 29, 2026Official source ↗
v0.7.13Oct 28, 2026Official source ↗
v0.7.12Oct 22, 2026Official source ↗
v0.7.11Oct 14, 2026Official source ↗
v0.7.10Sep 23, 2026Official source ↗
v0.7.9Sep 4, 2026Official source ↗
v0.7.8Aug 25, 2026Official source ↗

Quality profile

algorithm quality-v1.0 · assessed Oct 16, 2026

Quality Score92

Risk Score0

Confidence 1.00

Quality Score and Risk Score are different measures. A high score does not guarantee safety.

Score components

Maintenance · 2020

last activity 1 days ago

Documentation · 1515

README 22812 chars, description present

Ready to use · 1515

install instructions: yes; releases: yes

Metadata · 1010

4/4 of description, homepage, license, topics

Releases · 1010

10 release(s), latest 2 days ago

Category confidence · 106

classifier confidence 0.6

Adoption · 106

368 stars (log-dampened, 10k = max)

License clarity · 55

SPDX Apache-2.0

Community · 55

issues enabled, 35 forks

License

Apache-2.0

Known licensefrom the repository host

This is the catalog policy status, not legal advice.

Sources

github.com/fajarhide/omniPrimary★ 368 starslast activity Oct 14, 2026Primary repository · Confidence 1.00

Reviews

No reviews yet.

Report a problem with this card

Alternatives

Trail-CodeSimilar

Turbocharge Claude Code, Cursor, Codex, Gemini & every coding agent: faster, cheaper, with contextual understanding specific to your codebase.MIT · Developer Tools

graphifySimilar

Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local deterministic AST parsing, every edge explained, no vector store.Apache-2.0 · Developer Tools

ai-setupSimilar

Continuously sync your AI setups with one command. Codebase tailor suited agent skills, MCPs and config files for Claude Code, Cursor, and Codex.MIT · Developer Tools

ouroborosSimilar

Agent OS: the agent gets smarter on its own. We just hold the line: Interview-gated, staged evaluation, budgeted evolution loop. MCP server, 14 runtimes: Claude Code, Codex CLI, Gemini CLI, OpenCode, Copilot, Kiro and more.MIT · Developer Tools

lazycodexSimilar

The one and only agent harness for complex codebases. Project memory, planning, execution, and verified completion inside Codex.MIT · Developer Tools

README

From fajarhide/omni · license Apache-2.0 · fetched Oct 30, 2026

[CI] [Release] [License: Apache 2.0]

brew install fajarhide/tap/omni && contexere init

OMNI is now Contexere, Latin for "to weave together" and the root of "context". The omni command keeps working, so existing installs need no change. Contexere is growing into one place for the skills, credentials and memory every agent uses; today it ships the context and token layer.

Why

A coding agent reads everything a command prints: progress bars, layer hashes, a file it already read ten minutes ago. Every byte is billed and takes room the agent needs for the actual work.

Contexere takes out what the model does not need, before the model reads it. Nothing is deleted: whatever it removes stays on your machine, and one command brings it back.

Features

Reads a file the agent already read: sends one line with a handle instead of the file again.
Runs a noisy build or test: keeps the result and the failures, drops the progress noise.
Prints a token or password in output it rewrites: replaces it with [REDACTED] before the model sees it.
Fails a command: passes it through as printed, so the real error reaches the agent. A credential printed by a failing command reaches the agent too.
Reads JSON, YAML or CSV: leaves it byte for byte.
contexere doctor checks that each agent's hooks fire and redact.

It costs about 21 ms per command. Savings depend on what your agent does; run contexere stats for your own figure. Method and full numbers are on the benchmarks page.

Install

brew install fajarhide/tap/omni

Where	Command
macOS, Linux, WSL	`curl -fsSL omni.weekndlabs.com/install \| bash`
Windows (PowerShell)	`irm omni.weekndlabs.com/install.ps1 \| iex`

Existing installs keep working under omni, which is the same program.

Usage

contexere init      # pick your agents: Claude Code, Codex, Cursor, VS Code and more
contexere doctor    # check it is wired, or `contexere doctor --fix`
contexere stats     # what Contexere saved on your own history

After init, the hooks run on their own. There is nothing to prefix.

Supported agents

Tier	Agents
Full (rewrites tool output before the model reads it)	Claude Code, Codex CLI, Gemini CLI, OpenClaw, Hermes, Pi, Aider (via `cmd \| contexere`)
Handoff-first (shortens what the agent runs through `omni_run`)	Cursor, Windsurf
MCP-only (memory and recall, plus `omni_run`)	Cline, Roo, OpenCode, VS Code, Zed, Copilot, Antigravity

Codex runs only hooks it trusts: after contexere init --codex, start codex once and approve them (#359).

Documentation

Contributing

Issues and pull requests are welcome. Start with CONTRIBUTING.md.

Security

Do not open a public issue. Report privately; details in SECURITY.md.

License

Apache License 2.0.

Open the repository ↗