How does LockLLM pricing work?

LockLLM uses pay-as-you-go pricing. You only pay when built-in features activate - $0.0001 for threat detection, policy violations, or PII detection, and 5% of cost savings for smart routing. Prompt compression starts free with TOON. Proxy API usage is variable or free with BYOK.

What are tier rewards?

Tier rewards are free monthly credits you earn automatically based on your previous month's spending. The more you use LockLLM, the higher your tier and the more free credits you receive each month (ranging from $0.50 to $1000).

LockLLM is for individuals, developers, and teams who want safer AI. Scan prompts and files in the dashboard, or use the API to protect LLM apps and AI agents from prompt injection, jailbreaks, and risky instructions.

How do I use LockLLM for production traffic?

Route requests through the LockLLM API using an API key to get a risk signal before prompts reach your model or tools.

What is BYOK (Bring Your Own Key)?

BYOK allows you to use your own API keys from AI providers like OpenAI, Anthropic, and others. With BYOK, you only pay LockLLM for detection and routing fees, not for the actual AI API usage.

Do I need to retrain or fine-tune anything?

No. LockLLM works as a security gateway and does not require retraining, fine-tuning, or prompt rewrites.

Do you offer free credits?

Yes. All new users receive free credits when they sign up. Additionally, active users automatically earn free monthly credits based on their usage tier, ranging from $0.50 to $1000 per month. Safe scans are always free, and you only pay when threats are detected.

Will my credits expire?

Purchased credits never expire and remain in your account until used. Free tier credits earned through monthly rewards are provided as ongoing benefits as long as you maintain your tier level through active usage.

Does LockLLM include PII detection?

Yes. LockLLM can detect sensitive data like names, emails, phone numbers, and more. You can configure it to warn, block, or automatically redact PII from your requests. A $0.0001 fee applies when PII is detected, the same as other detection fees.

Does LockLLM support prompt compression?

Yes. LockLLM offers three compression modes to reduce token usage and lower inference costs. TOON is free and compresses JSON data (30-60% savings). Compact costs $0.0001 per use and compresses any text (30-70% savings). Combined costs $0.0001 and applies both sequentially for maximum compression. Configure via the X-LockLLM-Compression header.

Am I charged for every scan?

No. You are only charged when LockLLM actually detects a threat, policy violation, or PII in your request. If the scan result is safe, there is no charge at all. This means the vast majority of your requests cost nothing.

Simple pay-as-you-go pricing

Pay Only When We Add Value

Unlimited scanning with no upfront costs. Safe scans are always free. Only pay when we detect threats, policy violations, or optimizes your AI costs. Plus, earn free monthly credits as you grow.

Pricing

Safe scans are always free. Only pay when threats are detected.

Pay-As-You-Go

$0.0001/ flag

Pay only for inference or when threats are detected. Free credits included.

Get Started ->

BYOK

All AI Models

Use your API keys for inference. Pay only detection fees. Free credits included.

Configure ->

Enterprise

Custom

Dedicated support, SLA, and custom integrations.

Contact Sales ->

Pricing Details

Safe scan results

Free Safe scan results

Threat detection fee

$0.0001 Threat detection fee

Custom Threat detection fee

Policy violation fee

$0.0001 Policy violation fee

Custom Policy violation fee

PII detection fee

$0.0001 PII detection fee

Custom PII detection fee

Smart routing fee

5% of savings Smart routing fee

Custom Smart routing fee

Prompt compression

Free - $0.0001 Prompt compression

Custom Prompt compression

Proxy API usage

Variable cost Proxy API usage

Free (use your keys) Proxy API usage

Custom Proxy API usage

Platform

Production-ready API

Included Production-ready API

Clear allow/flag decisions

Included Clear allow/flag decisions

API key protection

Included API key protection

Debug timings in responses

Included Debug timings in responses

Smart routing, caching & compression

Included Smart routing, caching & compression

Dedicated support

24/7 priority Dedicated support

SLA guarantees

99.9% uptime SLA guarantees

Custom integrations

Available Custom integrations

Tier rewards

Up to $1000/month Tier rewards

Custom bonuses Tier rewards

Rate limits

Up to 200,000 RPM Rate limits

Custom limits Rate limits

Supported AI providers

200+ models Supported AI providers

All providers Supported AI providers

All + custom Supported AI providers

Threat Detection

Prompt injection

Included Prompt injection

Jailbreaks & policy bypass attempts

Included Jailbreaks & policy bypass attempts

Roleplay manipulation

Included Roleplay manipulation

Included Roleplay manipulation (“ignore rules” prompts)

Instruction override attempts

Included Instruction override attempts

Included Instruction override attempts (“ignore previous”)

System prompt extraction / secret leakage attempts

Included System prompt extraction / secret leakage attempts

Tool / function-call abuse (agent hijacking)

Included Tool / function-call abuse (agent hijacking)

RAG / document injection (poisoned context)

Included RAG / document injection (poisoned context)

Indirect injection (webpages, emails, PDFs)

Included Indirect injection (webpages, emails, PDFs)

Obfuscated / encoded attacks (evasion techniques)

Included Obfuscated / encoded attacks (evasion techniques)

Multi-vector prompt attacks (combined techniques)

Included Multi-vector prompt attacks (combined techniques)

Policy Protection

Custom content policies

Unlimited Custom content policies

Built-in safety categories

Included Built-in safety categories

Real-time policy violation detection

Included Real-time policy violation detection

Enhanced Real-time policy violation detection

Configurable enforcement (allow/warn/block)

Included Configurable enforcement

Advanced Configurable enforcement

Policy violation analytics & reporting

Basic Policy violation analytics

Advanced Policy violation analytics

Traffic Protection

Bot-generated content detection

Included Bot-generated content detection

Enhanced Bot-generated content detection

Excessive repetition detection

Included Excessive repetition detection

Enhanced Excessive repetition detection

Resource exhaustion protection

Included Resource exhaustion protection

Enhanced Resource exhaustion protection

Burst pattern detection

Included Burst pattern detection

Enhanced Burst pattern detection

Duplicate request filtering

Included Duplicate request filtering

Enhanced Duplicate request filtering

Pattern-based abuse scoring

Basic Pattern-based abuse scoring

Advanced Pattern-based abuse scoring

Data Protection

PII detection

Included PII detection

Automatic PII redaction

Included Automatic PII redaction

Configurable PII actions

Included Configurable PII actions

PII analytics in activity logs

Basic PII analytics

Enhanced PII analytics

Support

Email Support

Dedicated Support Support

Tier Rewards System

Earn rewards automatically based on your monthly usage. Higher tiers unlock increased rate limits and free monthly credits.

Tier	Name	Monthly Spend	Rate Limit	Monthly Reward
1	Starter	Free	300 RPM	Signup Bonus
2	Bronze	$10+	500 RPM	$0.50
3	Silver	$50+	1,000 RPM	$2.00
4	Gold	$100+	2,000 RPM	$5.00
5	Platinum	$250+	5,000 RPM	$15.00
6	Diamond	$500+	10,000 RPM	$40.00
7	Emerald	$1,000+	20,000 RPM	$80.00
8	Ruby	$3,000+	50,000 RPM	$250.00
9	Sapphire	$5,000+	100,000 RPM	$500.00
10	ObsidianHighest	$10,000+	200,000 RPM	$1000.00

Tiers are automatically assigned on the 1st of each month based on your previous month's spending

Stop overpaying for AI security

LockLLM helps individuals and teams secure LLM applications without heavyweight platforms or long-term contracts.

AI security rings illustration for LockLLM prompt scanning

Built for teams shipping real AI products, not experiments.
Detect prompt injection, jailbreaks, and instruction overrides before execution.
Scan prompts manually in the dashboard or enforce checks via API keys.
Clear, consistent detection results that are easy to interpret.
Freemium pricing with usage-based rewards and free credits.
Actively maintained as new attack patterns emerge.

What teams say about LockLLM

From individual developers to small teams, LockLLM is used to secure LLM applications without slowing development.

Mike Smets

We added LockLLM in front of our LLM endpoint and immediately caught prompts we wouldn’t have noticed in review.

Mike Bryan

The dashboard is great for quick checks, but the API is what really helped us enforce guardrails in production.

Liz Park

It’s hard to reason about prompt risks at scale. LockLLM gave us a simple signal we could actually act on.

Jessie Maison

We didn’t want to retrain models or build custom rules. Dropping in LockLLM was the fastest path to baseline safety.

Devani Janssen

Other security tools felt heavy. LockLLM stays out of the way and does one job well.

Mark Gerkules

The manual scanner makes it easy to check suspicious prompts without wiring anything into our system.

Andrey Propenk

We use the API key flow to protect multiple environments without changing our application code.

Jay Master

The results are straightforward to interpret. We could quickly tell when a prompt needed attention without digging into logs.

Steve Wolf

LockLLM helped us treat prompt security as infrastructure, not an afterthought.

Getting started with LockLLM

Everything you need to know

What is LockLLM?: LockLLM is an AI security and optimization API for LLM applications. It comes with built-in threat detection, custom content policies, smart routing, prompt compression, response caching, PII redaction, and abuse protection - all through a single integration.
Do I need to retrain my model?: No. LockLLM works independently of your AI model/application and does not require fine-tuning, retraining, or changes to your prompts at all whatsoever.
Do you have a free tier?: Yes. Safe scans are always free. You earn free monthly credits based on usage. Only pay small fees when threats are detected or when smart routing saves you money.
Does LockLLM offer free credits?: Yes. You get free credits when you sign up. Active users also automatically earn additional free monthly credits based on their usage levels.
Is this meant for developers or non-technical users?: Both. Developers can integrate LockLLM via API, while non-technical users can use the dashboard to scan and review prompts manually.
How do I use LockLLM?: You can scan prompts manually in the dashboard, or protect live traffic by routing requests through the LockLLM API using an API key.
Can I control or customize detection behavior?: Yes. You can create custom content policies to extend the built-in safety categories, configure scan actions (allow/block/warn), enable smart routing for cost optimization, enable PII detection with configurable actions (warn/block/strip), and adjust sensitivity per request. Chunking is available for long documents.
What should I scan besides user prompts?: Scan retrieved context (RAG snippets, webpages, knowledge-base text), file content after extraction, and tool inputs and outputs. These are common sources of indirect prompt injection.
Are there rate limits?: Yes. Limits are set high and are mostly there to prevent abuse and DDoS. It rarely affects normal usage. In the event you exceed your limit, you will receive HTTP 429 with a Retry-After header. Wait for the specified time, then retry.
What happens when a scan is unsafe?: Treat unsafe as a control point. Common actions are blocking, asking the user to rewrite, stripping instructions and keeping only facts, or routing to a restricted mode that cannot call tools.
Do you store my scan text?: LockLLM processes your text to return a result. We do not store, log, or train on any of your text input.
Where should LockLLM sit in an agent pipeline?: Place it at boundaries: before the model call, before tool execution, and when bringing in external context like RAG or web content. This helps prevent agent hijacking and tool abuse.
Can I use LockLLM for files and documents?: Yes. Extract text from PDFs, docs, logs, or tickets, then scan it the same way as prompts. For long documents, enable chunking for deeper coverage.
What are custom content policies?: Custom policies let you define your own content rules to prevent inappropriate AI output. Enforce content moderation guidelines alongside core injection detection to control what your AI generates.
How does smart routing work?: Smart routing analyzes task type and prompt complexity to automatically select the optimal model. Simple tasks route to efficient models, complex tasks to advanced ones. You pay 5% of cost savings when routing to cheaper models.
What is prompt compression?: Prompt compression reduces token usage to lower inference costs. TOON mode is free and compresses JSON data (30-60% savings). Compact mode compresses any text for $0.0001 per use (30-70% savings). Combined mode applies both sequentially for maximum compression.
What is BYOK and how does it work?: BYOK (Bring Your Own Key) lets you use your own API keys for all major AI providers. You only pay LockLLM for scanning and routing fees, while AI usage goes through your own keys. This gives you full cost control.
How do I earn free credits?: You receive free credits when you sign up. Active users also automatically earn additional free monthly credits based on their usage. Credits and rate limits scale with activity, rewarding consistent users with more benefits.
Can my team share a single account?: Yes. LockLLM supports organizations where teams share credit balances, custom policies, routing rules, and API keys. Create an organization from your dashboard to get started.
Do credits expire?: Purchased credits never expire and remain in your account until used. Free tier credits earned through monthly rewards are provided as ongoing benefits as long as you maintain your tier level through active usage.

Guardrails for every prompt

Secure your AI with confidence

Scan prompts manually in the dashboard, or protect live traffic with API keys that enforce safety checks in real time.

Try for free->