perch is live · pricing from $0.10/1K rule · last verified 6 Nov 2026 · viability healthy
perch
Semantic code linting: write custom plain-English rules for every pull request and catch AI-written bugs and security gaps.
Perch earns a real look when your review friction is project-specific behavior — the "refunds process once" or "no public buckets" rule no off-the-shelf linter knows about. The loop of turning an agent's mistake into a committed rule is the idea worth testing, and with a $5 free credit and cached scans free, a trial costs almost nothing. The tradeoff is on you: you author the rules, and results are confidence-scored findings rather than a single pass/fail verdict.
Verdict generatedVerified 7d ago · liveness 71/100 · cite: rightaichoice.com/tools/perch
- Teams whose review friction is project-specific behavior — money in cents, UTC timestamps, single-process refunds — not style.
- Engineering leads gating PRs written by Claude Code, Codex or Cursor agents that keep repeating the same mistakes.
- Polyglot codebases (Python, TypeScript, Go, Rust, Java, PHP and more) needing one rule file across languages.
- Small teams that want policy committed and code-reviewed next to the source in perch.yaml.
- Teams wanting zero-config drop-in linting — the distinctive value depends on you writing perch.yaml rules.
- Organizations needing a hosted dashboard for multi-user policy management or org-wide analytics.
- GitLab-first teams that need hosted PR review — Perch Cloud reviews GitHub pull requests, with the CLI as the GitLab path.
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Perch if you want zero-config drop-in linting that works without writing any rules of your own — Perch's value comes from rules you author in `perch.yaml` and scope with a glob, not from a shipped ruleset you switch on.
Every rule you want enforced costs you authoring time first — Perch ships built-in defect and security scans, but `no-hardcoded-keys`-style policy checks only exist once you write them in `perch.yaml`.
Perch prices per check rather than per seat: 10¢ per 1,000 rule checks, about 50¢ for a mid-sized pull request, with the first $5 of usage free and cached checks on unchanged code costing nothing. That puts it well below the AI review tools Perch itself benchmarks — Macroscope at roughly $0.95/PR, Greptile at roughly $1/PR on $30/seat, CodeRabbit at $1–3/PR on $24–60/dev/mo, Claude Code Review at $15–25/PR — and makes it cheap for small teams and solo developers who'd balk at a per-seat
In short
perch — Semantic code linting: write custom plain-English rules for every pull request and catch AI-written bugs and security gaps. Best for Teams whose review friction is project-specific behavior — money in cents, UTC timestamps, single-process refunds — not style., Engineering leads gating PRs written by Claude Code, Codex or Cursor agents that keep repeating the same mistakes., Polyglot codebases (Python, TypeScript, Go, Rust, Java, PHP and more) needing one rule file across languages.. Plans from $0.10/1K rule.
What is perch?
perch — Semantic code linting: write custom plain-English rules for every pull request and catch AI-written bugs and security gaps. Best for Teams whose review friction is project-specific behavior — money in cents, UTC timestamps, single-process refunds — not style., Engineering leads gating PRs written by Claude Code, Codex or Cursor agents that keep repeating the same mistakes., Polyglot codebases (Python, TypeScript, Go, Rust, Java, PHP and more) needing one rule file across languages.. Plans from $0.10/1K rule.
Perch is a semantic linter built for code review. Instead of a fixed rulebook, you write the rules yourself in plain English — "store timestamps in UTC," "refunds can only process once," "S3 buckets are never public" — and Perch enforces them on every changed method in a pull request. Rules live in a perch.yaml file beside your source, so they get reviewed and versioned like code. The engine parses your repo with tree-sitter and maps each method's callers and callees, so a rule is checked with surrounding context rather than as an isolated string match. Rules can scope to a whole project, a single file, or every method matched by a glob like infra/*.tf or api/**/*.py. From the terminal you can add a rule, test one against a single symbol, or filter a scan to a named rule. Unchanged code is cached and answered free. With no rules written, Perch still scans every changed method for defects and security issues: off-by-one loops, unchecked empty values, weak hashing, missing authorization. It supports Python, TypeScript, JavaScript, Go, Rust, Java, Kotlin, C, C++, C#, Swift, Ruby, PHP and six more. The CLI installs via npm under an MIT license; Perch Cloud reviews GitHub pull requests, and the CLI runs in GitLab CI. A Perch skill targets Claude Code, Codex, Cursor and Pi GitHub — turn a correction into a rule, check it against the broken code, then check the fix. Against Macroscope, Greptile, CodeRabbit and Claude Code Review, the vendor's published table puts Perch at about 20¢ per pull request. It fits teams whose review friction is project-specific behavior, not formatting, and it costs you the rule authoring.
Key features
- Semantic code review on every changed method in a pull request
- Author custom rules in perch.yaml with plain-English ensure guarantees
- Map each method's callers and callees using tree-sitter parsing
- Scope rules to a whole project, a single file, or methods matched by a glob
- Add rules from the terminal with perch rules add <name>
- Test a rule against one symbol with perch check <file>::<method>
- Built-in scans for off-by-one loops, unchecked empty values, weak hashing, missing authorization
- Confidence-scored findings with PR comments and failing checks
- Cached answers for unchanged code, free on repeat scans
- Fail CI jobs on breaking findings
- Perch skill for Claude Code, Codex, Cursor and Pi GitHub
- CLI distributed via npm under MIT license and runs in GitLab CI
- Perch Cloud reviews GitHub pull requests
- Supports Python, TypeScript, JavaScript, Go, Rust, Java, Kotlin, C, C++, C#, Swift, Ruby, PHP and six more
- Findings deleted 30 days after each scan; vendor states it does not train on your code
Behind the verdict
The pitch is narrow and honest: Perch is a semantic linter for the rules your team keeps repeating in review. If your pull requests keep tripping on UTC timestamps, integer cents, or single-process refunds, that is exactly the gap a fixed rulebook leaves open, and Perch lets you close it with a sentence in perch.yaml. Pick it when the friction is project-specific and recurring. The strongest fit is a team running coding agents that pass broken fixes into review — the 2026-09-30 post walks through a swallowed error and a rule that fails the PR is the kind of thing Perch is built to catch. The skill for Claude Code, Codex and Cursor makes the correction-to-rule loop short. Pass when nobody will own the rule file after the trial. The distinctive value depends on you writing and maintaining rules; with zero rules you get the built-in defect and security scans, which are useful but not the reason to choose it. GitLab-first teams also need to know Perch Cloud reviews GitHub pull requests, with the CLI as the GitLab path. In practice the confidence-scored findings take getting used to if you expect a binary verdict. Cached answers mean repeat scans are free and stable, which helps when you are tuning a rule against real code. We would start with two or three rules tied to mistakes that already cost you review time, then add more as they prove themselves. On cost, the vendor's published comparison against Macroscope, Greptile, CodeRabbit and Claude Code Review puts a typical 60-method pull request at about 20¢, with 10¢ per 1,000 rule checks and the first $5 free. Treat that as the vendor's estimate and validate against your own PR sizes. Where it bites: rule authoring is a real skill, polyglot repos need rules that hold across languages, and teams wanting a hosted
Models under the hood
as of 2026-11-13
Limitations
- Perch is distributed as an npm CLI (`npm install -g @lakeday/perch`) and custom rules are written in `perch.yaml` using plain-English `ensure` statements.
- Billing is usage-based (10¢ per 1,000 rule checks, ~20¢ per typical pull request, first $5 free), with cached/unchanged code free.
- Perch Cloud reviews GitHub pull requests and the CLI can run in GitLab CI, surfacing findings as PR comments and failing checks rather than an interactive review session.
as of 2026-10-13
Researching perch? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Is perch still active?
perch is live · pricing from $0.10/1K rule · last verified 6 Nov 2026 · viability healthy.
Viability score
How well maintained and how widely used is perch? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: November 2026
How we score →What's new in perch
Checked 11 days agoAcross the latest 6 updates: 1 launch and 5 news mentions.
Perch benchmarked on Greptile's 50 pull requests and ReviewBench
Perch published results on Greptile's 50 pull requests and ReviewBench, covering recorded bug detection, precision, thresholds, model costs, and reproduction scripts.
Perch rule targets AI-generated PRs that pass broken fixes into review
Post walks through a real swallowed error and a plain-English Perch rule that fails a pull request, aimed at coding agents that cheat on tests or report fixes that didn't happen.
Perch benchmarks Jev and other decision models on code fixes and lint rules
Perch compared Jev against other decision models for finding bugs and security vulnerabilities in code fixes and semantic lint rules.
Using Perch to check AI-generated landing page habits
Write-up describes using Perch to check the landing-page habits the author had been asking Codex to fix.
Perch demonstrates semantic linting with Jev
Example showing how Perch applies a semantic lint rule using the Jev decision model.
Perch launches as a semantic linter for bugs and security
Introducing Perch: a semantic linter for bugs, security issues, and project-specific rules, distributed with a CLI under the MIT license.
What do users say about perch?
What people actually say about perch — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
ResearchedDiscussed by users across 6 sources (Reddit, Hacker News, YouTube, Product Hunt, GitHub, Lemmy) · researched Sep 26, 2026.
Weighted by the 70 posts each of 6 sources contributed.
- +Free and globally installed via npm as @lakeday/perch — zero cost to try.
- +Natural-language rules in perch.yaml let you encode team conventions as reviewable code.
- +Findings include rule ID, file:line, severity, confidence percentage, and enclosing method.
- +Per-finding confidence scoring avoids binary pass/fail and reduces false-positive fatigue.
- +Rules can target non-code files, like enforcing sentence case in Markdown docs.
- −Almost no independent user reviews or testimonials exist for the actual linter.
- −Open language-filtering bug means shipped scans may flag irrelevant vulnerability classes.
- −No listed integrations with CI/CD, editors, or Slack — everything is manual CLI wiring.
- −Name collision with Google Perch and a board game makes community support hard to find.
- −Unclear how the Jev engine performs at scale or on large monorepos.
- • No monetary hidden costs since the tool is free, but the time cost of writing and maintaining perch.yaml rules is real
- • If Perch Cloud launches, it may introduce paid tiers for cloud features not currently available
- • Manual CI/CD integration effort if you want to run perch scan outside your local machine
How we verified this
Every listed fact about perch — pricing, features, links, status — is re-read from the vendor's own pages by our verification pipeline and each change is logged with a date; the last pass ran on 6 Nov 2026. Analysis is drafted with AI models and checked against our editorial standards before publishing; see how we verify.
Verification history
We have re-verified perch 6 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — changed platforms, limitations
- — changed who should skip it, description, summary, our verdict, who it suits, features, pricing tiers
- — changed limitations, pricing page
- — changed who it suits, summary, description, our verdict, features, pricing tiers
- — changed our verdict, integrations, description, pricing tiers, who should skip it, who it suits, summary, features
- — changed SSL certificate
Free to cite with attribution — this page re-verifies continuously.
How much does perch cost in 2026?
perch: Plans from $0.10/1K rule. Pricing last verified 6 Nov 2026.
Plans compared
Every published perch tier, its price as the vendor publishes it, who it fits, and what it adds over the tier below. The 12-month cost projection follows.
| Plan | Price | Billing | Who it's for |
|---|---|---|---|
| Free credit | $5Vendor: “$5 free credit” | — | — |
| Perch Cloud (usage-based) | $0.10/1K ruleVendor: “10¢ per 1,000 rule checks” | usage | — |
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Where the pricing makes sense
The company stage and team size where perch's pricing actually pencils out — and where peers do it cheaper.
Perch prices per check rather than per seat: 10¢ per 1,000 rule checks, about 50¢ for a mid-sized pull request, with the first $5 of usage free and cached checks on unchanged code costing nothing. That puts it well below the AI review tools Perch itself benchmarks — Macroscope at roughly $0.95/PR, Greptile at roughly $1/PR on $30/seat, CodeRabbit at $1–3/PR on $24–60/dev/mo, Claude Code Review at $15–25/PR — and makes it cheap for small teams and solo developers who'd balk at a per-seat
Has perch's pricing changed?
No plan price for perch has moved since we began logging in Aug 2026; the pricing page is re-read on every verification pass.
Who should pick perch?
Pick perch if you are Teams whose review friction is project-specific behavior — money in cents, UTC timestamps, single-process refunds — not style., Engineering leads gating PRs written by Claude Code, Codex or Cursor agents that keep repeating the same mistakes., Polyglot codebases (Python, TypeScript, Go, Rust, Java, PHP and more) needing one rule file across languages., Small teams that want policy committed and code-reviewed next to the source in perch.yaml..
Look elsewhere if you are Teams wanting zero-config drop-in linting — the distinctive value depends on you writing perch.yaml rules., Organizations needing a hosted dashboard for multi-user policy management or org-wide analytics., GitLab-first teams that need hosted PR review — Perch Cloud reviews GitHub pull requests, with the CLI as the GitLab path., Codebases where nobody will own or maintain the rule file after the trial run..
Skip it if Skip Perch if you want zero-config drop-in linting that works without writing any rules of your own — Perch's value comes from rules you author in `perch.yaml` and scope with a glob, not from a shipped ruleset you switch on.
GeneratedReal-world workflow fit
Concrete scenarios for the personas perch actually fits — and what changes day-one when you adopt it.
A Codex or Claude Code agent opens a PR that passes tests but discounts orders by 150%. You turn on Perch Cloud scans for the repository in the dashboard; the PR gets a Perch Scan check and a line-level comment, and you add `discount-limit --where cart.py --each method --ensure "Discounts can be no larger than 20%."` from the terminal.
Outcome: A later PR that reintroduces the oversized discount fails the Perch Scan check automatically, and unchanged code returns the same cached answer on every rerun instead of you re-arguing the review.
You write `record-ownership` scoped to `api/**/*.py` update methods so ownership is checked before a record is modified, test it against one symbol with `perch check src/transfer.py::transfer --rules record-ownership`, then wire `perch scan` into CI and fail the job on breaking findings with exit code 3.
Outcome: Ownership violations block the merge at the CI gate rather than surfacing as a review comment a tired reviewer might scroll past.
You opt into the documentation checks and target `docs/**/*.md` with rules like `sentence-case`, `docs-headings-name-things`, `docs-no-anthropomorphism`, and `docs-show-output` to require fenced command blocks to show real output.
Outcome: Doc drift gets caught per pull request inside the free $5 of usage, since caching means unchanged doc files cost nothing on subsequent runs.
Use cases
- Add a `private-logs` rule scoping `src/**/*.ts` methods so passwords and access tokens never reach log statements.
- Enforce a `discount-limit` business invariant on `cart.py` so discount methods never exceed 20%.
- Lint Markdown documentation for sentence-case headings by targeting `docs/**/*.md` files.
- Require `atomic-writes` in Python methods so transactions roll back when any write fails.
- Gate AI-generated pull requests by scanning agent-authored code for defects and weak cryptography before merge.
- Enforce `record-ownership` checks in `api/**/*.py` update methods to prevent unauthorized record modification.
- Catch vague narrative headings and filler prose in docs with `docs-headings-name-things` and `docs-no-anthropomorphism`.
- Require that every fenced command block in your docs shows its real output via the `docs-show-output` rule.
Setup time & first value
How long it actually takes to get something useful out of perch — broken out by persona, not the marketing-page minute.
Solo developer: minutes — `npm install -g @lakeday/perch` then `perch scan`, since built-in defect and security scans run without configuration. Small team: plan on an afternoon for the first genuinely useful week — you install the CLI, install the Perch skill in whichever of Claude Code, Codex, or Cursor your agents use, write one or two `perch.yaml` rules, and turn on repository scans in the
Switching to or from perch
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From a comment-only AI review tool: keep it and add Perch alongside, since Perch's output is a committed rule that reruns rather than a one-time review comment.
- →From Semgrep: run Perch side by side during a trial so your custom Semgrep rules stay enforced while you re-express the behavior-critical ones as plain-English `ensure` statements.
- →From manual review checklists: convert each recurring correction into a `perch rules add <name> --where <glob> --each method --ensure` command and let CI enforce it.
- →From a CI workflow that shells out to a linter: replace the workflow with a Perch Cloud pull-request scan turned on for the repository in the dashboard, which needs no workflow file.
- ↗To Semgrep: reimplement your `perch.yaml` rules as Semgrep patterns if you decide you need a mature, community-authored rule ecosystem more than natural-language intent.
- ↗To CodeRabbit or Greptile: move to those if your team would rather receive AI review comments on a PR than maintain committed policy in your repository.
- ↗To a bespoke CI linter: export the behavior statements you care about into your own script if you need binary pass/fail gating instead of confidence percentages.
- ↗To no linter at all: because rules live in `perch.yaml` next to your source, deleting the file and the CLI removes Perch cleanly with nothing hosted to unwind.
Integrations
Resources & guides
Tutorials & Learning
YouTube returned 6 videos for “perch”, and we withheld 6: 6 could not be judged, because “perch” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about perch.
Official links
What are the best perch alternatives?
The closest alternatives to perch we list are CodeRabbit, Greptile and Codacy AI. Ranked by direct product-type match, not category overlap.
| Alternative | Pricing | Starts at | Last verified |
|---|---|---|---|
| CodeRabbit | Paid | Pricing not published | — |
| Greptile | Freemium | Pricing not published | — |
| Codacy AI | Freemium | Pricing not published | — |
Alternatives to perch
View allCodeRabbit
AI code review that validates, prioritizes, and secures every pull request your team ships.
Tools that pair well with perch
Common stack mates teams adopt alongside perch, with the specific reason each pairing earns its keep.
perch vs its closest rival
perch's closest head-to-head is with Amazon Codewhisperer — These aren't competitors — they're different layers of the stack. If you want an AI assistant that writes and modernizes code inside your IDE with AWS-native security scanning and agentic tasks, pick Amazon Q Developer (the current name for CodeWhisperer since April 2024). If you want a lightweight, free CLI that enforces your team's own rules — logging secrets, discount caps, ownership checks — across AI-generated code from Cursor, Codex, or Claude Code, pick Perch. The realistic answer for a team running AI coding agents: use both. Q Developer generates, Perch gates. Verdict generated
Head-to-head comparisons
Perch vs Amazon Codewhisperer
These aren't competitors — they're different layers of the stack. If you want an AI assistant that writes and modernizes code inside your IDE with AWS-native security scanning and agentic tasks, pick Amazon Q Developer (the current name for CodeWhisperer since April 2024). If you want a lightweight, free CLI that enforces your team's own rules — logging secrets, discount caps, ownership checks — across AI-generated code from Cursor, Codex, or Claude Code, pick Perch. The realistic answer for a team running AI coding agents: use both. Q Developer generates, Perch gates.
GeneratedPerch vs Snyk Deepcode Ai
These are different purchases despite both being code security tools. Snyk DeepCode AI is a platform commitment: you get hybrid symbolic+ML detection, Agent Fix autofixes, context-aware prioritization, and a security dashboard — but enterprise-grade features like Evo AI pentesting and coding-agent security sit behind an Enterprise Platform Subscription with credit-based pricing that scales by active contributors. Perch costs nothing and does one job well: it enforces your team's own semantic rules with probabilistic confidence scores, and it fits teams gating AI-agent output in Claude Code, Codex, or Cursor. If nobody on the team will write perch.yaml, Snyk is the safer pick. If budget is zero and the real problem is project-specific behavior — logging secrets, discount caps, ownership checks — Perch wins on cost and on rules that live in your repo.
GeneratedPerch vs Endor Labs
These aren't really competitors, so treat it as a 'what problem are you buying for' question rather than a bake-off. Endor Labs is for security and DevSecOps teams that need reachability-verified vulnerability prioritization plus governance over AI coding agents, MCP servers, and skills — with FedRAMP 2026's reachability mandate (Aug 2026 news) pushing that capability from nice-to-have to requirement. Perch is for a small Python/TypeScript team that wants its own repo-committed natural-language rules — logging secrets, discount caps, ownership checks — enforced in CI for free. If you have a dedicated security function and compliance obligations, pick Endor Labs; if your review friction is project-specific behavior and you'll write the rules yourself, pick Perch.
GeneratedPerch vs Mindgard
These two do not compete for the same budget, so there is no head-to-head winner. If your problem is adversarial risk in production AI agents — shadow AI, guardrail bypasses, exploit-backed findings your GRC team can audit — Mindgard is built for exactly that, and you will pay a contact-sales enterprise price for it. If your problem is code-level policy enforcement inside a repo (off-by-one loops, MD5-hashed tokens, unhandled nulls, or catching what Claude Code/Codex/Cursor wrote), Perch is free, runs from your terminal or GitHub Actions, and expects you to write your own perch.yaml rules. Pick by problem, not by category: agent attack surface vs source-tree defects.
GeneratedPerch vs Bito
These are two different line items on the same engineering budget, and most teams adopting agents will eventually want both. Bito is the one that pays for itself only at scale — it needs multi-repo agent traffic already burning tokens, and it comes with a scoping call instead of a published rate. Perch is free and installs in minutes, so its real cost is the time you spend writing perch.yaml rules; it earns its place the moment your review friction is project-specific behavior rather than style. If you can only pick one right now: pick Perch to stop bad agent-generated code from merging, and pick Bito once your agent bill is big enough that a routing and grounding layer has something to save.
GeneratedPerch vs Coro
These two will never appear on the same shortlist. If you're a lean IT team or an MSP trying to collapse endpoint, email, cloud, identity, network and data security into one agent and one console, Coro is built for exactly that — but budget for a partner-issued quote and accept it isn't aimed at a dedicated SOC doing deep threat hunting. If you write Python or TypeScript and your review friction is project-specific behavior rather than style, Perch is free and worth an afternoon: install it, write one perch.yaml rule, and see whether probabilistic findings with confidence scores beat your current linter. Don't evaluate them against each other; evaluate each against its own alternative.
GeneratedFrequently asked questions
For makers of perch
- Get a Verified Listing for perch ($39): a written decision within 1 working day.
- Free listing for another tool: Free reviews are decided in order, within 3 weeks of the review starting.
Used perch? Help shape our editorial sentiment research.