A secure external API for FaithGPT responses with clean Markdown, Scripture-aware guardrails, documented JSON shapes, and request-level usage accounting.
Base URL
https://api.faithgpt.io/v1
Auth
Bearer API key
Contract
JSON + OpenAPI
Base URL
https://api.faithgpt.io/v1Endpoints
4
Guardrails
10
Artifacts
11
Protocol
HTTPS
Sign-in required
Signed-in Max customers can create, pause, resume, revoke, and inspect API key prefixes from the dashboard view. Raw keys are shown once.
API key
Sign in to create keys
Requests
Sign in
Recent API request records
Credits Used
Sign in
Backend-calculated usage
Completed
Sign in
Successful persisted requests
Average Latency
Sign in
Measured from real requests
Daily API Usage
Last 7 UTC days
Day 1
Day 2
Day 3
Day 4
Day 5
Day 6
Day 7
Usage appears here after real API requests are recorded for this account.
Rate Limit Status
Endpoint Breakdown
No endpoint usage recorded yet.
Recent Requests
View docs| Timestamp | Endpoint | Status | Credits | Latency |
|---|---|---|---|---|
| Real request records will appear here after this account calls the API. | ||||
Sign in with a Max account to create and manage API keys.
Create a key to view real prefixes, status, scopes, caps, and usage controls. Raw keys are shown only immediately after creation.
Authenticated request
curl https://api.faithgpt.io/v1/faithgpt/answer \
-H "Authorization: Bearer fgpt_live_your_key" \
-H "Content-Type: application/json" \
-d '{
"prompt": "Explain Romans 8:28 with pastoral care.",
"responseFormat": "markdown",
"maxOutputTokens": 600
}'Every runtime URL uses https://api.faithgpt.io/v1. Use your own key in the Authorization header.
Real request builder
Sign in and enable API access before creating keys. Without a key, the playground can still build requests but cannot run them.
API keys are still validated by the backend. This page never stores the key you enter.
curl https://api.faithgpt.io/v1/faithgpt/answer \
-H "Authorization: Bearer fgpt_live_your_key" \
-H "Content-Type: application/json" \
-d '{
"prompt": "Explain Romans 8:28 with pastoral care.",
"responseFormat": "markdown",
"maxOutputTokens": 600
}'Quick links
Endpoint docs are generated from /api/faithgpt-api.docs.json. Benchmark harness routes stay out of the public OpenAPI contract.
Return a clean Markdown FaithGPT answer using the production harness without FaithGPT rendering tags.
{
"prompt": "Explain Romans 8:28 with pastoral care.",
"responseFormat": "markdown",
"maxOutputTokens": 600
}{
"id": "string",
"answerMarkdown": "string",
"citations": "array",
"usage": "input/output token details",
"billing": "backend-calculated credit usage metadata"
}Access
FaithGPT API keys are sent as Bearer tokens, stored as secure hashes, scoped by endpoint family, and can be rotated or revoked.
Authorization: Bearer fgpt_live_your_keyCreate and manage keys from your signed-in FaithGPT account. The raw key is shown once at creation; future views show only the prefix.
Hash API keys at rest; show only key prefix after creation.
Require scopes for each endpoint family.
Block API access unless the signed-in account has API access enabled.
Apply per-key, per-user, per-IP, per-route, and global rate limits.
Preflight estimated credit usage and stop requests above configured caps.
Reconcile credits after actual input and output tokens are known.
Strip FaithGPT rendering tags such as <bible> before external responses.
Separate instructions from untrusted project knowledge, files, and web sources.
Store request IDs, latency, route, model, status, token accounting, abuse flags, and safety decisions.
Never expose benchmark evaluation materials, runner metadata, provider keys, deployment URLs, or operations paths.
Backend calculated
Credit usage is calculated on the backend, estimated before generation, and reconciled after actual input and output tokens are known.
Account managed
Account and API controls can configure key scopes, daily and monthly usage caps, rate limits, revocation, and request telemetry.
Responses
X-RateLimit-LimitMaximum requests or credit units allowed in the current window.
X-RateLimit-RemainingRemaining capacity before the key is limited.
X-RateLimit-ResetUnix timestamp when the current window resets.
Retry-AfterSeconds to wait before retrying when a 429 response is returned.
JSON
{
"error": {
"message": "Too many requests. Try again in 12 seconds.",
"type": "rate_limit_error",
"code": "rate_limited"
},
"request_id": "req_..."
}bad_requestMalformed JSON or a missing required field.
invalid_api_keyMissing, malformed, revoked, or unknown API key.
insufficient_creditsThe request exceeds account credits or configured caps.
insufficient_scopeThe key does not have access to this endpoint.
rate_limitedThe key, account, IP, or route exceeded a configured limit.
server_errorUnexpected server error. Retry with exponential backoff.