Pricing

Pricing

Every plan includes the full scoring engine. Enterprise unlocks policy enforcement, audit trails, and config governance.

Free
$0
forever
  • 10 lifetime optimizations
  • 5 per minute rate limit
  • All 19 tools
  • All 3 output formats
  • Unlimited scoring & checking
Start Free
Power
₹899
per month (~$11)
  • Unlimited optimizations
  • 60 per minute rate limit
  • All 19 tools
  • Always-on mode
  • Priority support
Get Power
Enterprise
Custom
contact us
  • Unlimited optimizations
  • 120 per minute rate limit
  • Policy enforcement (advisory + enforce modes)
  • Tamper-evident audit trail (JSONL, hash-chained)
  • Config lock with passphrase protection
  • Custom rules engine (up to 25 rules)
  • Session lifecycle management
  • Session retention policies (auto-purge)
  • Dedicated support
Contact Sales See all Enterprise features

Full Feature Comparison

Every tool, every tier. All 19 tools are available on all plans — metering applies only to optimization calls.

Feature Free Pro Power Enterprise
Limits & Rate
Monthly optimizations 10 lifetime 100 / mo Unlimited Unlimited
Rate limit 5 / min 30 / min 60 / min 120 / min
Always-on mode
Analysis Tools
optimize_prompt metered
check_prompt free
classify_task free
Routing Tools
route_model free
pre_flight metered
Refinement Tools
refine_prompt metered
approve_prompt free
Cost & Compression Tools
estimate_cost free
compress_context free
prune_tools free
Configuration Tools
configure_optimizer free
get_usage free
prompt_stats free
License Tools
set_license free
license_status free
Session Tools
list_sessions free
export_session free
delete_session free
purge_sessions free
Enterprise-Only Governance
Policy Enforcement
Audit Logging
Config Lock
Custom Rules
Session Retention Policies

Enterprise Features

The governance layer you need before approving AI in production.

🛡 Policy Enforcement

Switch from advisory mode (flags issues) to enforce mode (blocks prompts that fail policy). BLOCKING rules — both built-in and custom — gate every optimization. Risk threshold gating blocks high-risk approvals above a configurable score. Deterministic: same input produces the same verdict every time.

How to enable
configure_optimizer({ mode: "enforce", threshold: 60 })
Example output when blocked
{ "status": "blocked", "reason": "risk_score_exceeded", "risk_score": 72, "threshold": 60, "blocking_rules": ["hallucination_risk", "agent_underspec"], "recommendation": "Add source context and safety constraints" }

📜 Audit Logging

Every action generates a JSONL audit entry with SHA-256 hash chaining. If any line is deleted or modified, the chain breaks and tampering is detectable. Local-only, opt-in, and never stores raw prompt content. Answers the question: who changed what, when, and was it allowed?

How to enable
configure_optimizer({ audit_logging: true })
Example audit entry
{ "timestamp": "2026-03-01T10:15:32.441Z", "action": "optimize_prompt", "request_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890", "tier": "enterprise", "risk_score": 34, "verdict": "allowed", "prev_hash": "8f14e45f...abc123", "hash": "7c222fb2...def456" }

🔒 Config Lock

Lock your governance configuration with a passphrase. Once locked, no one can change policy, strictness, threshold, or audit settings without the correct secret. Every attempt — successful or blocked — is audit-logged. Tamper-evident: the lock state itself is hash-protected.

How to enable
configure_optimizer({ lock: true, passphrase: "your-secret" })
What happens on unauthorized change
{ "status": "error", "error": "config_locked", "message": "Configuration is locked. Provide the correct passphrase to modify settings.", "audit_logged": true }

📋 Custom Rules

Define your own regex-based rules in a JSON file. Each rule specifies a pattern, severity (BLOCKING or NON-BLOCKING), a risk dimension (hallucination, constraint, underspec, or scope), and a risk weight. Up to 25 custom rules. Rules are hashed for reproducibility and included in session exports.

Example custom rule
{ "id": "no_pii_in_prompt", "description": "Flag prompts that include email addresses", "pattern": "[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}", "applies_to": "all", "severity": "BLOCKING", "risk_dimension": "constraint", "risk_weight": 20 }
Validate with CLI
prompt-lint --validate-custom-rules

🗑 Session Retention Policies

Auto-purge sessions older than a configurable retention period. Dry-run mode previews what would be deleted before committing. The keep_last parameter protects your newest N sessions from purge. Only session files are affected — config, audit logs, and license data are never deleted.

How to use
purge_sessions({ older_than_days: 90, keep_last: 5, dry_run: true })
Example dry-run output
{ "dry_run": true, "sessions_matched": 12, "sessions_protected": 5, "sessions_to_delete": 7, "oldest": "2025-11-15T08:22:00Z", "newest_deleted": "2025-12-01T14:30:00Z" }

Built for Enterprise Trust

Addressing the concerns your security and compliance teams will ask about.

Security

  • Zero LLM calls inside — prompts never leave your machine
  • No cloud services, no telemetry, no phone-home
  • Runs entirely on localhost via MCP stdio transport
  • Ed25519 offline license validation — no network needed
  • Config lock with SHA-256 hashed passphrase protection

Privacy

  • All data stored locally in ~/.prompt-control-plane/
  • No user accounts, no email collection, no PII stored
  • License keys contain only: tier, dates, license_id
  • Audit log never stores raw prompts or compiled outputs
  • Works behind corporate VPNs and air-gapped networks

Compliance & Governance

  • Deterministic — same input produces same output, always
  • Tamper-evident audit trail with hash-chained JSONL entries
  • Policy enforcement blocks non-compliant prompts at approval
  • Reproducible exports with rule-set hashes for audit
  • Source-available — full code review available on GitHub

Deployment

  • No infrastructure to manage — runs as a local process
  • Compatible with Claude Code, Cursor, Windsurf, and any MCP client
  • No database, no containers, no cloud accounts required
  • Single npm install — works on macOS, Linux, and Windows
  • Zero dependencies on external AI services

Frequently Asked Questions

Common questions about plans, billing, and enterprise features.

What counts as an "optimization"?
Only three tools are metered: optimize_prompt, refine_prompt, and pre_flight. Everything else — scoring, checking, routing, cost estimation, compression, pruning, session management — is unlimited on all tiers.
Do I need an account to use the free tier?
No. The free tier works out of the box with zero configuration. No account, no API key, no credit card. Just install the MCP server and start using it.
How does the license key work?
License keys use Ed25519 asymmetric signatures and are validated entirely offline. No phone-home, no account server. The key contains your tier and expiry date, cryptographically signed. Activate it once with set_license and it works until expiry.
Can I switch between tiers?
Yes. Tier priority is: license key (cryptographically verified) > PROMPT_CONTROL_PLANE_PRO environment variable > default free. Upgrading is instant — activate a new license key and your limits update immediately.
What happens when I hit the free tier limit?
Metered tools return a clear error with your current usage, the limit, and purchase URLs for Pro and Power tiers. Non-metered tools continue working without restriction.
Is enterprise self-hosted or cloud?
Everything runs locally. There is no cloud component. The MCP server runs on your machine, all data stays on your filesystem, and audit logs are local JSONL files. Enterprise just unlocks governance features — it does not change the deployment model.

Ready to get started?

Start free with 10 optimizations. Upgrade anytime with an offline license key.