The belief layer for AI agents.
Instantly align your agents with your own worldview.
A structured, queryable REST API that gives AI agents access to a user's beliefs, worldview, values, and faith-based preferences — so agents can make decisions aligned with who that person actually is.
The Problem
AI agents are making decisions on behalf of humans — without understanding what those humans believe, value, or practice.
An agent booking travel doesn't know about Shabbat. It optimizes for price, not for arriving before sundown on Friday.
An agent ordering groceries doesn't know about halal restrictions. It picks the cheapest option, not the permissible one.
An agent managing investments doesn't know about riba prohibition. It maximizes returns without regard for ethical screens.
An agent making ethical trade-offs doesn't know if you're a deontologist or utilitarian. It has no moral compass to reference.
Guardrails fills this gap.
The 7 Domains
Every Guardrails profile organizes a person's values across seven structured domains — queryable independently or together.
Prayer schedules, holy days, dietary laws, tithing rules, and sacred time boundaries.
Ethical orientation, moral foundations scores, and decision-making principles.
Heritage, language preferences, cultural celebrations, and community obligations.
Veganism, minimalism, sustainability pledges, and daily routines.
Interest policies, investing screens, charitable giving, and debt philosophy.
Advance directives, blood transfusion preferences, organ donation, and care priorities.
Social issue stances, policy priorities, activism, and civic engagement.
API Reference
Simple, predictable REST endpoints. JSON in, JSON out. Authenticated with bearer tokens and user consent.
| Method | Endpoint | Description | Cost |
|---|---|---|---|
| GET | /v1/profiles/{user_id} | Get Full Profile | $0.05 |
| GET | /v1/profiles/{user_id}/domains/{domain} | Query Single Domain | $0.01 |
| GET | /v1/profiles/{user_id}/check | Compatibility Check | $0.005 |
| POST | /v1/profiles/{user_id}/decide | Decision Support | $0.10 at base tier |
| GET | /v1/profiles/{user_id}/calendar | Faith-Aware Calendar | $0.02 |
| POST | /v1/profiles | Create Profile | $0.005 |
| GET | /v1/taxonomies/{domain} | Get Taxonomies | $0.005 |
| POST | /v1/config/{type} | Create Configuration | Free |
| GET | /v1/config/{type} | List Configurations | Free |
| GET | /v1/config/{type}/{name} | Get Configuration | Free |
| PUT | /v1/config/{type}/{name} | Update Configuration | Free |
| DELETE | /v1/config/{type}/{name} | Delete Configuration | Free |
Configurations API
Go beyond the built-in 7 domains. Create custom value domains, define specialized rule logic, and save reusable presets — all scoped to your API key.
Define new value domains beyond the built-in 7. Create industry-specific taxonomies like halal_supply_chain or environmental_impact and reference them in checks and decisions.
Custom check logic, domain weights, and prompt templates. Define rules like "block if dietary_laws.system == halal" with keyword lists and weighted scoring across domains.
Named combinations of domains, rulesets, and response formats. Apply any preset via ?preset=your-preset on standard endpoints for instant, reusable configurations.
Interactive Playground
Click "Run Request" to hit the live Guardrails API. These are real responses from api.useguardrails.com — not mocks.
Request
/v1/profiles/usr_demo/check
Body
Response
Request
/v1/profiles/usr_demo/check
Body
Response
Request
/v1/profiles/usr_demo/calendar?days=30
Response
Request
/v1/profiles/usr_demo
Response
Request
/v1/config/preset
Body
Response
Universal Integration
Guardrails is a standard REST API with a comprehensive OpenAPI specification. Any AI agent — regardless of platform — can use it.
OpenAI / GPT
Import our OpenAPI spec as function definitions. GPT agents call Guardrails endpoints via tool-use.
Anthropic / Claude
Define Guardrails endpoints as tools. Claude's tool-use calls our REST API directly.
Salesforce Agentforce
Register Guardrails as an External Service via our OpenAPI spec. Available to any Agentforce topic.
Any Agent
If your agent can make HTTP requests, it can use Guardrails. No SDK required. No protocol lock-in.
OpenAPI Specification
A well-documented OpenAPI 3.1 spec that any agent platform can consume directly.
openapi: 3.1.0
info:
title: Guardrails API
version: 1.0.0
description: The belief layer for AI agents.
servers:
- url: https://api.useguardrails.com/v1
paths:
/profiles/{user_id}:
get:
operationId: getProfile
summary: Get a user's full Guardrails profile
/profiles/{user_id}/domains/{domain}:
get:
operationId: getDomain
summary: Query a single domain
/profiles/{user_id}/check:
get:
operationId: checkCompatibility
summary: Quick compatibility check
/profiles/{user_id}/decide:
post:
operationId: getDecision
summary: Values-aligned decision support
/profiles/{user_id}/calendar:
get:
operationId: getCalendar
summary: Faith-aware calendar
/profiles:
post:
operationId: createProfile
summary: Create a new profile
/taxonomies/{domain}:
get:
operationId: getTaxonomy
summary: Get domain optionsPricing
Every API call is metered. The more you use, the less you pay per call. No monthly fees, no commitments.
| Monthly API Calls | Per-Call Rate | Estimated Monthly Cost |
|---|---|---|
| 1 – 10,000 | $0.02 | $0.02 – $200 |
| 10,001 – 100,000 | $0.015 | $150 – $1,500 |
| 100,001 – 500,000 | $0.01 | $1,000 – $5,000 |
| 500,001 – 1,000,000 | $0.007 | $3,500 – $7,000 |
| 1,000,001+ | $0.005 | Contact us |
Every API call is billed at the rate for your current monthly volume tier. As your usage grows within a billing cycle, your per-call rate automatically decreases.
The /decide endpoint uses AI reasoning and costs 5x the standard per-call rate for your tier (e.g., $0.10 at the base tier).
Need custom volume pricing, SLAs, on-prem deployment, or dedicated support? contact@useguardrails.com
MCP Integration
Connect Guardrails directly to Claude, Cursor, Windsurf, or any MCP-compatible agent. No HTTP calls to write — your agent gets 11 tools out of the box.
Drop this into your agent's MCP settings — that's it.
{
"mcpServers": {
"guardrails": {
"url": "https://mcp.useguardrails.com/mcp"
}
}
}Every tool accepts your api_key as a parameter. Your agent handles this automatically once configured.
It can now check values, make aligned decisions, query calendars, and manage profiles — all through natural tool calls.
checkRule-based compatibility checkdecideLLM-powered decision (BYOK)calendarFaith-aware schedulingcreate_profileCreate user values profileget_profileRetrieve profile dataupdate_profileUpdate profile domainsget_taxonomiesList value domainslist_configurationsList custom configscreate_configurationCreate custom configupdate_configurationUpdate custom configdelete_configurationDelete custom configCompatible with
Get Started
Enter your email. We'll send you your API key and a verification link. No passwords, no credit card — just start building.
Already have an account? Get a login link
Explore Guardrails
Deep-dive into how Guardrails works across 7 value domains, 7 API endpoints, and 12 real-world use cases.
Prayer schedules, holy days, dietary laws, tithing rules, and sacred time boundaries.
Ethical orientation, moral foundations scores, and decision-making principles.
Heritage, language preferences, cultural celebrations, and community obligations.
Veganism, minimalism, sustainability pledges, and daily routines.
Interest policies, investing screens, charitable giving, and debt philosophy.
Advance directives, blood transfusion preferences, organ donation, and care priorities.
Social issue stances, policy priorities, activism, and civic engagement.
Retrieve a user's complete Guardrails profile across all permitted domains.
Retrieve data from one specific domain of a user's profile.
Quick yes/no check if a product, service, or action is compatible with a user's values.
Pass a scenario with options and get a values-aligned recommendation.
Get a user's upcoming religious observances, blocked times, and holy days.
Create a new Guardrails profile for a user.
Retrieve the structured taxonomy of options for a specific domain.
Create a custom domain, ruleset, or preset scoped to your API key.
List all configurations of a given type for your API key.
Retrieve a specific configuration by name.
Update an existing configuration's body or description.
Permanently remove a configuration.
An AI grocery agent that automatically filters products by halal certification.
A travel agent that respects Sabbath boundaries when booking flights and hotels.
A financial agent that screens investments against the user's ethical filters.
A healthcare agent that honors patient values in advance directive decisions.
A meal-planning agent that respects dietary philosophy and sustainability goals.
A calendar agent that blocks prayer times and religious observances automatically.
An AI assistant that adapts tone, greetings, and formality to cultural norms.
A content moderation agent that applies the user's personal moral standards.
A giving agent that schedules donations aligned with faith and financial commitments.
A shopping agent that enforces sustainability pledges on every purchase.
An event-planning agent that respects every attendee's dietary and scheduling constraints.
A news agent that curates content based on the user's actual values, not engagement metrics.