Guardrail by NeatProxy

The runtime firewall
for AI coding agents.

See coding-agent spend. Check your budget before the next supported request reaches the provider.

Install freeExplore spend scenario ↓

Blocking requires a supported proxy integration. Codex with a ChatGPT subscription is visibility only; Cursor support is planned. Check compatibility.

Local proxy · No prompt storage · Mode-specific enforcement

Scenario inputs stay in memory across marketing pages and audience switches. Reloading or entering the console clears them; shared links contain only the audience.

The problem

Agents spend money faster than anyone can watch.

A coding agent runs unattended for minutes at a time, and a single retry loop can burn more than a month of subscription before anyone looks. The tools that exist today tell you afterwards.

01

Bills arrive late

Provider billing is a monthly rear-view mirror. By the time a number looks wrong, the money is already spent.

02

Counters only count

A token counter watches the meter run. It has no opinion about when to stop, and no way to act on one.

03

Alerts are not brakes

A threshold alert fires after the spend it is warning you about. A budget you cannot enforce is a preference.

Inspect a real analysis of fictional data

A total tells you what. A sample shows you why.

Try the bundled sample audit, inspect the contributing rows, and read the rule behind each finding. No upload or email required.

Analyze a sampleRead evidence & limitations

Spend scenario

Explore the numbers. See the assumptions.

Model a possible opportunity, or compare fictional requests with and without a cap. Neither is a savings guarantee.

Illustrative · no account connected

Your assumptions, made explicit. Nothing here changes a real budget.

Headcount changes the per-person breakdown, not total spend.

Enter $0–$1,000,000. Manual entry supports values above the slider range.

An assumption you choose, not measured waste or guaranteed recovery.

Free has no plan cost. Pro is priced per developer; Team is one bundle. Taxes excluded.

Choose assumptions, then calculate. Editing a value clears any previous result.

# pricing_model

pay for the cloud, not the firewall.

Guardrail runs free on your machine, with no credit card and no time limit. Pro adds the hosted dashboard, longer history, and sync across every machine you own. Team is one flat plan for the whole group.

[ developer - local first ]
$0 / forever

// free

Local firewall, budget policies, local dashboard, 7 days of history. No credit card required.

[ get installation script ]
Most Popular
[ developer - cloud sync ]
$15 / per month

// pro

Everything in Free, plus the hosted dashboard, 90 days of history and sync across machines. $144 yearly.

[ upgrade to pro ]
[ team - cloud sync ]
$99 / month

// team (5 seats)

A flat bundle for up to 5 seats, with an admin console, per-member spend and shared project caps.

[ start teams ]
[ compare full feature matrix & FAQ ]

// TWO BILLING MODELS. ONE REAL RISK.

Your login mode determines what Guardrail can control.

Blocking requires a supported proxy integration. Codex with a ChatGPT subscription is visibility only; Cursor support is planned.

FLAT SUBSCRIPTION SEATS(Subscription logins)

Visibility is not the same as blocking

Subscription usage and token-equivalent costs are not your provider bill. What Guardrail can see and control depends on the tool and how you sign in.

What this integration can do:
  • •Claude Code requests routed through the local proxy can be checked against budget policies.
  • •Codex with a ChatGPT subscription provides usage telemetry only. It cannot block requests or guarantee remaining subscription quota.
USAGE API SPEND(Supported Anthropic / OpenAI proxy modes)

Check the policy before sending

Metered requests can add cost while an agent retries. A supported proxy integration checks the configured policy before forwarding the next request.

What this integration can do:
  • •Pre-send hard dollar caps on localhost:4000 before calls hit the wire.
  • •A blocked request returns a policy error. How the agent retries or recovers depends on the tool.

How it works

One command in front of your agent.

For supported proxy integrations, Guardrail sits between your coding tool and the provider. It forwards requests and records usage metadata, so policies can block the next call before it is sent.

1

Connect your tool

Connect Claude Code with your existing login or Codex with an API key. Codex subscription mode provides telemetry, not proxy enforcement.

2

It reads the metadata

Supported proxied requests flow through localhost:4000 byte-for-byte. Guardrail records model, tokens, cache and estimated cost — not prompt text, responses, or code.

3

See it, then cap it

A live local dashboard shows spend, sessions and hidden cost. Set a budget and an over-budget call is blocked before it ever reaches the provider.

Claude Code · Codex (API key)
──▶
Guardrail · :4000
──▶
Anthropic · OpenAI
└──▶
SQLite · metadata only

The request body is forwarded untouched. Only model, token counts and cost estimates are written locally to 127.0.0.1

What Guardrail controls

Spend control, not another dashboard to babysit.

Local visibility and control for supported integrations. Cloud sync is a separate feature. See how each one works.

Control

Budgets that block

Caps on dollars, requests, tokens or requests-per-minute. Over budget, the call is never sent, so it costs nothing.

Visibility

Hidden cost, surfaced

The cache-write tax on a first turn and the reasoning tokens you never see, attributed per request.

Privacy

Metadata stays local by default

Personal usage metadata stays on your machine unless you enable sync. Team sync defaults are explained in the privacy guide.

Coverage

Support depends on your login

Claude Code and Codex API-key requests support blocking. Codex subscription usage is visibility only; agent frameworks are in preview.

Forecast

Know the month before it lands

Spend pace and a projected month-end total from the sessions you have already run.

Flow

Pause without restarting

Flip tracking off and back on from the dashboard. Your tool keeps working throughout.

Why Guardrail is different

Prevention, not reporting.

Most tools in this space observe. Guardrail sits in the request path, which is the only place a budget can actually be enforced.

Provider dashboards

Show you the bill

Accurate and far too late. They are scoped to an account, not to the project or the session that caused the spike, and they cannot stop anything.

Token counters

Show you the meter

Useful for curiosity, limited for control. Watching a number climb does not stop the next request from being sent.

Observability platforms

Analyse after the fact

Rich analysis, but it arrives after the spend, and it usually means sending prompts and responses to a third party to store.

Guardrail

Stops the request

Runs on your machine, in the path, with your own credentials. Policies are evaluated before each call, so an over-budget request is never sent and never billed.

FAQ

The questions everyone asks first.

Do you store my prompts or code?
No. Guardrail records metadata only— model, token counts, cost, timings, status. Prompt text, responses, code and credentials never touch its logs or database; a privacy layer and disk-scan tests enforce it.
Do I have to give Guardrail an API key?
It depends on the integration. Claude Code can relay your existing login; Codex proxy enforcement uses your OpenAI API key. Codex with a ChatGPT subscription uses telemetry instead and cannot block requests. See the compatibility matrix.
Can it actually block spend, or just watch it?
On supported proxy integrations, budget policies are checked before the request leaves your machine. An over-budget or denied-model call is stopped, no upstream call is made, and your tool gets a provider-shaped policy error. Recovery depends on the tool. Blocking requires a supported proxy integration. Codex with a ChatGPT subscription is visibility only; Cursor support is planned.
Which tools does it cover?
Blocking requires a supported proxy integration. Codex with a ChatGPT subscription is visibility only; Cursor support is planned. Agent frameworks using the OpenAI-compatible Chat Completions API are in private preview. See the tool and login-mode matrix.
Is the cost my real provider bill?
It is an estimatefrom a configurable price table, shown with confidence labels — a close guide, not your invoice. Reconcile against provider billing for the exact number.
What happens if I stop Guardrail?
Start it again with guardrail start and your tools carry on. While the daemon is stopped it is still in the request path, so bring it back up before running an agent; a fail-open relay that passes traffic through untracked is on the roadmap.
Where does my data live, and what about Windows?
In a local SQLite file under ~/.guardrail, bound to localhost. Personal cloud sync is opt-in; Team machines sync metadata by default on supported versions and can opt out. It runs on macOS, Linux, and Windows through WSL2.
Is there a free plan?
Yes, and it does not expire. Free gives you the local proxy, budget policies, blocking, and the local dashboard with 7 days of history, with no credit card required. The local Free plan needs no email or activation code. Request a code only to try Pro cloud features. Paid plans add the hosted dashboard, 90 day history, and sync across machines.
Is pricing per developer or per workspace?
Pro is per person ($15 per month, or $144 per year). Team is a flat bundle, not a per-seat rate: one price covers up to 5 or 10 seats, so adding a teammate inside your bundle costs nothing extra.

Put a runtime budget around your AI agents.

get guardrailview pricingjoin our discord