AI coding agents spend most of their input tokens on tool output: build logs, test runs, git status, the same file read again after every edit. All of it goes into the context window raw, and it is re-sent on every turn after that.
sqz compresses tool output before it enters the context. Per-command formatters keep what an agent acts on (the failing test, the assertion, the file:line) and drop what it does not. Anything the model already has in context comes back as a short reference instead of the content. Source code passes through unchanged, and so do stack traces, migrations and credential lines.
cargo install sqz-cli sqz-mcp # or: brew install ojuschugh1/sqz/sqz · npm i -g sqz-cli · pipx install sqz
sqz init # hooks + MCP server + agent guidance for every client it finds
Real output from the release binary (assets/demo.tape). Regenerate with vhs assets/demo.tape.
Without sqz: With sqz:
cat auth.py (200 lines): 1,300 tokens 1,300 tokens (source is served in full)
cargo test (1 failure): 366 tokens 185 tokens (failure, assertion, file:line)
sed -n '41,80p' auth.py: 260 tokens 35 tokens (§ref:…:L41-80§ to the read above)
git status, unchanged: 166 tokens 37 tokens (§ref:…§)
───────────────────────────────────── ─────────────
Total: 2,092 tokens 1,557 tokens (26% saved, nothing dropped)
Single Rust binary, deterministic, no LLM calls, works offline. Every compressed result can be recovered byte-exact with sqz expand, except output that mentions credentials, a stack trace or a migration, which sqz never caches.
Note 注意
Name disambiguation: this repo, ojuschugh1/sqz, is an independent project and is not affiliated with any other similarly named or working tool or other compression projects that shorten “squeeze”. If you installed sqz / sqz-cli / sqz-mcp from crates.io, npm, PyPI, or Homebrew, it comes from this repository.
Token Savings
24.7% average reduction across 3,003 real compressions ·
13-token refs for output the model already has ·
100% of agent-critical facts kept in the quality benchmark ·
0% on source code, and on stack traces and credential lines
One developer’s week, measured from actual sqz gain output:
Single-command compression (measured via cargo test -p sqz-engine benchmarks):
Content
Before
After
Saved
Repeated log lines
148
39
74%
Large JSON array
259
111
57%
JSON API response
64
53
17%
Git diff
61
54
11%
Prose/docs
124
122
2%
Stack trace (safe mode)
82
82
0%
Session-level with dedup
Anything the model already has in its context comes back as a reference. Measured with the release binary against a throwaway database, using the fixtures in sqz/tests/quality_bench.rs and demo/, in cl100k tokens of everything the agent sees (the reference plus sqz’s one-line note on stderr):
Repeat
Without sqz
With sqz
Saved
Unchanged git status run again
166
37
78%
Lines 41-80 of a 200-line file read earlier
260
35
87%
Same cargo test output, nothing changed
366
39
89%
A word on what actually repeats. A user who counted 92 of their own Claude Code sessions (script) found identical whole-file re-reads in 1 of 542 reads, and line-range re-reads in 8.4% of them. So sqz does not lead with “the same file read five times”: the repeats that happen in practice are unchanged command output, line ranges of a file already read, and files re-read after a small edit, and each has its own reference type (details). Sessions with noisy command output and repeated commands see the biggest wins.
Install
Prebuilt binaries (no compiler required — works on every platform):
# macOS / Linux
curl -fsSL https://raw.githubusercontent.com/ojuschugh1/sqz/main/install.sh | sh
# Windows (PowerShell)
irm https://raw.githubusercontent.com/ojuschugh1/sqz/main/install.ps1 | iex
# Any platform via npm
npm install -g sqz-cli
# macOS / Linux via Homebrew
brew tap ojuschugh1/sqz
brew install sqz
Build from source via Cargo:
cargo install sqz-cli sqz-mcp
sqz-cli provides the sqz binary; sqz-mcp provides the MCP server. sqz-engine is a library dependency — it compiles automatically and does not need to be installed separately.
Build from source (cargo install sqz-cli) works too, but needs a C toolchain:
Linux: build-essential (apt) or equivalent
macOS: Xcode Command Line Tools (xcode-select --install)
Windows: Visual Studio Build Tools with the “Desktop development with C++” workload. Without these, cargo install fails with linker link.exe not found. If you don’t already have them, use the PowerShell or npm install above instead.
Then initialize:
sqz init --global # hooks apply to every project on this machine
# or
sqz init # hooks apply to just this project (.claude/settings.local.json)
--global writes to ~/.claude/settings.json (the user scope per the
Anthropic scope table),
so the sqz hook fires in every Claude Code session on this machine. This is
the common case on first install. Your existing permissions, env,
statusLine, and unrelated hooks in ~/.claude/settings.json are
preserved — sqz merges its entries rather than overwriting.
The entries are its hooks and two allow rules, Bash(set -o pipefail) and
Bash(sqz compress --cmd *), which cover only the parts its rewrite adds.
sqz does not approve commands; your own permission rules still decide.
When Claude Code asks before a rewritten command, or denies it, its message
quotes the rewritten command, such as set -o pipefail; git tag v0 2>&1 for
git tag v0.
Plain sqz init (project scope) is useful when you want sqz active only
inside one repo. If ~/.claude/settings.json already has the sqz hooks, it
leaves them out of the project file, since they already run in every project.
After installing, sqz doctor checks the whole chain — binary, database,
shell hook, which clients are detected vs actually routed through sqz, and
whether anything was compressed recently — and prints the fix for each gap
it finds. sqz doctor --fix applies those fixes. After upgrading sqz, run it
once: it brings hooks and guidance an older version wrote up to date, removes
hook files current sqz no longer uses and a project copy of hooks that
~/.claude/settings.json already has, points hooks whose sqz binary is gone
at the running one, and notes any hook that still runs a different sqz
binary. sqz uninstall removes only sqz’s own entries, deletes a file only
when nothing else is left in it, and ends by listing the files it kept.
Without a terminal to answer its prompt it needs --yes, like sqz reset.
See what sqz would have saved, before installing any hook
If you already use Claude Code or Kiro, your transcripts are on disk. Replay
them through the sqz engine and get a counterfactual estimate on your own
sessions, computed locally with nothing uploaded. From the author’s last
week of Kiro IDE work, trimmed to the summary:
$ sqz discover --replay
sqz discover — counterfactual replay (last 7 days)
────────────────────────────────────────────────────────
Kiro IDE: 35 sessions, 155 sub-agent runs, advisory
outputs tokens in avoidable
shell 6,297 1,917,662 454,028 23.7%
file reads 4,880 7,854,071 188,239 2.4%
other MCP servers 1,547 6,717,440 4,234,128 63.0%
already through sqz 260 1,393,517 -
out of reach 4,172 1,282,285 -
Avoidable with no help from the agent: 0 tokens
Avoidable if the agent uses every route: 4,876,395 tokens, 25.4% of all tool output
Of that, 3,816,588 is the MCP proxy's lossy first pass.
Each output is replayed the way sqz could actually have reached it on that
client. Kiro can’t let sqz rewrite a tool call, so the first total is zero
there; on Claude Code the shell rows split into commands the hook rewrites
on its own and ones it skips. The second total assumes the agent pipes shell
output through sqz compress, reads files with the sqz MCP tools (lossless,
so only repeats count) and runs MCP servers behind sqz-mcp proxy, and the
report says how much of that is the proxy cutting long strings and arrays on
first pass. Web fetches, edits and sub-agent results are out of reach on every
client and are counted, not replayed.
It scans ~/.claude/projects and ~/.kiro/sessions (Kiro CLI and IDE), or
point it anywhere with --transcripts PATH. Each transcript replays against a
fresh throwaway cache, so dedup references never pretend to span sessions,
and the report says “estimate” because that’s what it is: these outputs were
not compressed at the time.
Only using one agent? Pass --only (or --skip) to limit which
configs are written:
sqz init --only opencode # just OpenCode, nothing else
sqz init --only opencode,codex # OpenCode and Codex
sqz init --skip cursor,windsurf # everything except Cursor and Windsurf
Accepted names: claude, cursor, windsurf, cline, gemini,
kiro, opencode, codex. Aliases (claude-code, gemini-cli, roo,
kiro-cli) also work. --only and --skip can’t be combined.
Manual installation (preserve comments in your config)
sqz init round-trips your config file through a JSON parser to merge
the sqz entry, which drops any comments in your opencode.jsonc (and
the analogous JSON-with-comments files other tools accept). If you’ve
commented your config carefully and want to keep them, install by hand
instead.
OpenCode — two steps:
Drop the plugin file in place. sqz prints the generated TS to
stdout so you don’t have to hand-write the path-escaping logic:
Add the MCP entry to your existing opencode.jsonc yourself.
Append this block inside the top-level mcp object (create the
mcp object if it doesn’t exist):
Comments in the rest of your file stay put. OpenCode auto-discovers
the plugin file; no plugin array entry needed (adding one causes
double-loading, see issue #10).
The same plugin file loads on OpenCode 1 and 2. OpenCode 2 does not
load the plugin sqz 2.0.0 and earlier wrote, so after upgrading sqz run
sqz doctor --fix (or the print-opencode-plugin line above again) to
rewrite it.
Other tools: Claude Code and Gemini CLI use plain JSON configs
without comment support, Cursor, Windsurf and Cline get markdown rules
files, and Codex gets a TOML entry in ~/.codex/config.toml that keeps
your comments, so the automated path is non-destructive there. Use
sqz init --only <tool> for those. An existing .gemini/settings.json
gets sqz’s hook merged in; if it has comments or doesn’t parse, sqz init
leaves it alone and prints the entry to add by hand.
Run sqz doctor afterwards to see what was configured for each client
and which integration tier it is on.
How It Works
On transparent clients (Claude Code, Gemini CLI, Copilot CLI, OpenCode) sqz installs a hook that rewrites shell commands so their output goes through sqz before it reaches the AI tool. On advisory clients the agent is told to pipe output through sqz compress or use the sqz MCP tools. The table below gives each client’s tier.
Claude → git status → [sqz hook rewrites] → compressed output
JSON: strips nulls and empty values, cuts strings over 500 characters, compact encoding, TOON format
Logs — collapses repeated lines
Test output — shows failures only (state-machine parsers for Rust, Go, Python, JS, JVM)
What doesn’t get compressed:
Source code: served in full on first read
Stack traces, SQL migrations, PEM blocks and credential lines: routed to safe mode and passed through unchanged at any size, and never cached; on the shell hook a per-command formatter may still summarize the rest of a test run around them
Your prompts and the AI’s responses — controlled by the AI tool, not sqz
Supported Tools
Tool
Integration
Setup
Claude Code
PreToolUse + PostToolUse hooks + MCP server (transparent)
sqz init
Cursor
.cursor/rules/sqz.mdc guidance (advisory); MCP server added by hand
sqz init
Windsurf
.windsurf/rules/sqz.md guidance (advisory); MCP server added by hand
sqz init
Cline
.clinerules guidance (advisory); MCP server added by hand
sqz init
Gemini CLI
BeforeTool hook (transparent)
sqz init
Kiro
Steering + MCP server (advisory)
sqz init
OpenCode
TypeScript plugin for OpenCode 1 and 2 + MCP server (transparent)
Transparent clients have a hook that routes shell commands through sqz on
its own. On advisory clients sqz saves tokens only when the agent pipes
output through sqz compress or calls the sqz MCP tools; the client’s
built-in shell and file tools bypass it. sqz doctor shows each client’s
tier.
Compress Any MCP Server
sqz-mcp proxy sits between your agent and any stdio MCP server and compresses
what flows back: tool results go through the full sqz pipeline (dedup refs on
repeats, safe-mode for stack traces and secrets, error results untouched), and
verbose tool descriptions get compacted so tools/list stops eating your
context window. The proxy injects an sqz_expand tool so the agent can recover
any cached original byte-exact; results that mention credentials, a stack trace or
a migration are never cached.
Wrap a server by prefixing its command in your MCP config:
Flags: --lazy-tools shortens every tool description to one sentence and
injects an sqz_tool_help tool that serves the full original docs on demand
(some MCP servers spend 10-40k tokens on tools/list alone); --no-desc
keeps descriptions verbatim; --no-cache stores nothing, so it sends no
dedup refs and adds no sqz_expand tool. Works with
every MCP client (Claude Code, Cursor, Windsurf, Zed, Codex, Kiro, …)
because the client just sees a normal MCP server. Full guide, including the
server’s own sqz_read_file / sqz_grep / sqz_list_dir tools and per-client
config: MCP context compression.
CLI
sqz init --global # Install hooks for every project on this machine
sqz init # Install hooks for just this project
sqz init --only kiro # Only configure Kiro (skip the rest)
sqz init --only opencode # Only configure OpenCode (skip the rest)
sqz init --skip cursor # Configure every agent except Cursor
sqz init --ci # Also write .github/hooks/sqz.json for Copilot agent CI runs
sqz compress <text> # Compress (or pipe from stdin)
sqz compress --no-cache # Compress without dedup (always full output)
sqz expand <ref> # Recover original content from a §ref:HASH§ token
sqz compact # Evict stale context to free tokens
sqz reset # Clear dedup cache or compression stats
sqz gain # Show daily token savings (bar chart)
sqz gain --project . # Per-project daily gains
sqz gain --days 30 # Last 30 days
sqz stats # Cumulative compression report (includes regret signals)
sqz stats --cost # Estimated $ saved under a prompt-cached billing model
sqz stats --breakdown # Per-command token usage breakdown
sqz stats --project . # Stats for current project only
sqz stats --project list # List all tracked projects
sqz stats --share # Five-line share card sized for a paste into an issue or post
sqz doctor # Verify sqz is installed, wired into your clients, and active
sqz doctor --fix # Apply the fixes doctor prints (run once after upgrading)
sqz uninstall # Remove sqz's hooks, guidance and MCP entries, keep the rest
sqz discover # Find missed savings (incl. your weakest-compressing commands)
sqz discover --replay # Counterfactual replay of your agent transcripts (Claude Code, Kiro)
sqz recall "auth timeout" # Full-text search everything sqz has compressed
sqz resume # Re-inject session context after compaction
sqz vizit # Live terminal dashboard (like htop for AI agents)
sqz hook claude # Process a PreToolUse hook (Claude Code)
sqz hook claude-post # Process a PostToolUse hook (Claude Code)
sqz hook kiro # Legacy; Kiro now uses steering + MCP (sqz init)
sqz print-opencode-plugin # Print OpenCode plugin TS for manual install
sqz proxy --port 8080 # Local endpoint: compresses an API request body and sends it back, no forwarding
Dedup Escape Hatch
When sqz sees the same content twice, it returns a compact §ref:HASH§ token
instead of the full text. Most models handle this fine, but some (e.g., GLM 5.1)
can’t parse the ref format and loop. Four ways to work around this:
# 1. Recover original content from a ref
sqz expand a1b2c3d4 # prefix match
sqz expand '§ref:a1b2c3d4§' # paste the whole token
# 2. Compress without dedup (per-invocation)
echo "..." | sqz compress --no-cache
# 3. Disable dedup globally (env var)
export SQZ_NO_DEDUP=1
# 3b. Or just shorten the ref freshness window (seconds, default 1800).
# Useful on clients without a compaction hook; 0 disables refs.
export SQZ_REF_TTL_SECS=300
# 4. MCP passthrough tool (returns input byte-exact, zero transforms)
# Available via tools/list when sqz-mcp is running
Recall: search everything sqz has seen
Everything that flows through sqz is indexed locally (SQLite FTS5, on your
machine, nothing leaves it). When your agent compacts its context and loses
that error message from an hour ago, search for it instead of re-running
the command:
$ sqz recall "connection refused"
1. [git · 2h ago] … curl: (7) Failed to connect: >>connection refused<< on port 5432 …
full content: sqz expand 3f9a1c05e2d84b17
Agents get the same thing as the sqz_recall MCP tool.
Track Your Own Savings
Run sqz gain in your shell any time to see your own daily breakdown (see the
Token Savings section above for what the output looks like), and sqz stats
for the full cumulative report:
Add --breakdown to see exactly which commands consume the most tokens:
$ sqz stats --breakdown
🔍 Top Token Consumers
──────────────────────────────────────────────────────────────────────
command calls tokens in out saved
──────────────────────────────────────────────────────────────────────
dedup 249 45541 3237 93%
stdin 51 30851 24289 21%
auto 132 18288 7740 58%
echo 17 1050 558 47%
ls -la 8 948 948 0%
cargo build 7 170 145 15%
git status 4 56 8 86%
──────────────────────────────────────────────────────────────────────
sqz stats and sqz gain count cl100k tokens of what sqz actually printed,
abbreviation legend and expand hint included. A run of one character class
longer than 2 KB, such as a long separator line, is counted in 2 KB slices,
which can be off by about 1% plus 4 tokens.
Honest accounting: cost and regret
Token reduction is not automatically billed-cost reduction — with provider
prompt caching, most context re-transmits at a ~90% discount, and compression
that drops something the agent needed costs extra turns instead
(arXiv:2607.12161). sqz measures both
sides instead of hand-waving:
sqz stats --cost estimates dollars saved under a prompt-cached billing
model (cache write ×1.25 once, cache read ×0.10 per subsequent turn), with
every assumption printed and overridable (--price-in, --reread-turns,
--cache-write-mult, --cache-read-mult). This is the conservative
estimate: without caching the same savings would bill at full input price
on every turn.
sqz stats reports regret signals: quick re-runs (the agent re-produced
byte-identical output within 2 minutes — the repeat bought nothing; the
Claude Code PostToolUse hook never answers with a reference, so its
repeats count as ordinary compressions) and
ref expands (§ref§ tokens recovered to original bytes). Both are proxies
for “compression dropped something the model needed.” If one command
keeps showing up, its formatter needs work — file an issue.
sqz’s design already avoids the failure modes that study measured:
compression is deterministic and query-agnostic (never invalidates provider
prompt caches), only new tool output is touched (history is never rewritten),
commands that follow output, run a server or background a job are left alone
(on Claude Code outside bypass mode, the PreToolUse hook leaves pipelines,
loops and heredocs as written and the PostToolUse hook compresses their
output after a successful run), grouped lists and pipelines (Gemini CLI,
Copilot CLI, Claude Code in bypass mode) skip the per-command formatters,
and the 16-token net-win gate skips compressions that wouldn’t pay for their
own markers. A command is also left as written when its rewrite could run a
different set of commands than the original. In ls ${DEST:?}; echo next,
piping ls on its own would let echo next run after the failed expansion
stopped the shell, and where a list goes through sqz as one group, a failed
cd in cd build && make; make test would skip the group instead of only
make.
Per-project filtering:
sqz stats --project . # stats for current project only
sqz stats --project list # list all tracked projects
sqz gain --project . # daily gains for current project
sqz gain --days 30 # last 30 days instead of 7
sqz gain --days 30 --project . # combine both
Stats are stored locally in SQLite under ~/.sqz/sessions.db — nothing leaves your machine.
npm/pnpm/yarn/bun install/audit/outdated (test and run when they fail), tsc, eslint, biome
Python
pytest, ruff, mypy, pip
Go
go test (incl. -json stream), go build, go vet, golangci-lint
Cloud
terraform plan/apply/destroy/refresh/init
Containers
docker/podman ps/images, kubectl describe/apply
JVM
gradle build/test, maven
Mobile
xcodebuild (build + test), adb logcat
.NET
dotnet build/publish/test
System
grep/rg, tree, curl/wget
Other commands, including kubectl get and logs, gh, aws, gcloud, ls and find, go through the generic compression pipeline.
Table compactor — aligned-column output from tools without a dedicated formatter (ps aux, netstat, database CLIs) collapses its padding runs to two-space separators. The detector is strict — indented lines, code, YAML, and JSON never match.
Structural summaries: the engine has an API that reduces a code file to imports, function signatures and a call graph. The hook and MCP paths do not use it; they serve source in full.
Dedup cache: SHA-256 content hash, persistent across sessions. Byte-identical output the model already has = 13-token reference. A line range of a file the model already has in full (sed -n '40,80p', sqz_read_file with offset/limit) = §ref:HASH:L40-80§. A re-read after a small edit = the changed lines with one unchanged line on each side, through the sqz MCP tools (the shell hook sends the edited file again); edits spread through the file come back in full. References are only served while the original is still in context (30-minute window, tunable via SQZ_REF_TTL_SECS, reset on compaction).
JSON pipeline: strip nulls → drop bookkeeping fields (__v, etag, request and correlation ids) → flatten → collapse arrays → TOON encoding (lossless compact format). Fields whose value is an empty string, array or object are dropped at any depth, and strings over 500 characters are cut to their first 500 characters followed by ....
Safe mode: stack traces (a trace header followed by its frames), SQL DDL at the start of a line, PEM blocks and credential lines (key=value or key: value with a secret-looking key), with or without ANSI colors, skip the pipeline and pass through unchanged at any size, with only ANSI codes stripped, and are never cached. On the shell hook a per-command formatter still runs first. Its summary is used only when it keeps every line of each paragraph that holds one of them unchanged, so it may summarize the rest of a test run around a trace; otherwise the whole output passes through. Output that only mentions a password, key, trace or migration outside those structures is compressed as usual but is not cached either, so anything a lossy stage drops from it cannot be expanded.
If a stage panics, sqz compress prints the input unchanged with one line on stderr and keeps the command’s exit status, and the Claude Code PostToolUse hook leaves the result as it was.
Measured quality, including what each stage drops and what it never touches: quality benchmark. For the full technical details, see docs/.
Everything sqz persists (stats, dedup cache, sessions) lives in one SQLite
file, ~/.sqz/sessions.db by default. Set SQZ_DB_PATH to keep it
per-project instead:
# e.g. in the project's .envrc (direnv)
export SQZ_DB_PATH="$PWD/.sqz/sessions.db"
Every surface honors it — the shell hook, sqz stats/gain/expand, and
the MCP server (set it under env in your MCP config). Prefer absolute
paths: relative values resolve against whatever directory the process runs
from. The parent directory is created if missing (mode 0700, like ~/.sqz).
Privacy
Zero telemetry — no data transmitted, no crash reports
Fully offline — works in air-gapped environments
All processing local
Development
git clone https://github.com/ojuschugh1/sqz.git
cd sqz
cargo test --workspace
cargo build --release
If sqz saved you tokens today, a star helps other people find it. sqz stats --share prints your own numbers if you want to show them.
Contributors
Thanks to everyone who has contributed code, fixes, and ideas to sqz:
And to everyone who filed the detailed bug reports behind our fixes. Precise repros make this project better with every release. Want to join them? PRs are reviewed fast: see the open issues to get started.
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?v=4)
评论