systemprompt-api 0.14.6

Axum-based HTTP server and API gateway for systemprompt.io AI governance infrastructure. Exposes governed agents, MCP, A2A, and admin endpoints with rate limiting and RBAC.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155

use std::sync::Arc;

use axum::body::Body;
use axum::http::{HeaderValue, StatusCode};
use axum::response::Response;

use crate::services::gateway::audit::GatewayRequestContext;
use crate::services::gateway::protocol::inbound::InboundAdapter;
use crate::services::gateway::protocol::outbound::UpstreamError;
use crate::services::gateway::service::{DispatchInputs, GatewayService};

use super::RequestContext;
use super::extract::PreparedRequest;

pub(super) async fn dispatch_to_provider(
    rc: &RequestContext<'_>,
    inbound: Arc<dyn InboundAdapter>,
    prepared: PreparedRequest,
) -> Result<Response<Body>, (StatusCode, String)> {
    let PreparedRequest {
        principal,
        body_bytes,
        gateway_request,
        provider,
        upstream_model,
        session_id,
        context_id,
        gateway_conversation_id,
    } = prepared;

    let max_tokens = gateway_request.max_tokens;
    let is_streaming = gateway_request.stream;

    let gateway_ctx = GatewayRequestContext {
        ai_request_id: rc.ai_request_id.clone(),
        user_id: principal.user_id().clone(),
        session_id: Some(session_id),
        context_id,
        gateway_conversation_id: Some(gateway_conversation_id),
        trace_id: Some(principal.trace_id().clone()),
        provider,
        model: upstream_model,
        max_tokens: Some(max_tokens),
        is_streaming,
        wire_protocol: inbound.wire_name().to_owned(),
    };

    let gateway_config = rc
        .profile
        .gateway
        .as_ref()
        .and_then(systemprompt_models::profile::GatewayState::resolved)
        .ok_or_else(|| (StatusCode::NOT_FOUND, "Gateway not enabled".to_owned()))?;

    match GatewayService::dispatch(
        gateway_config,
        &rc.profile.providers,
        rc.ctx.db_pool(),
        DispatchInputs {
            request: gateway_request,
            raw_body: body_bytes,
            ctx: gateway_ctx,
            inbound,
        },
    )
    .await
    {
        Ok(resp) => Ok(resp),
        Err(e) => map_dispatch_error(&e),
    }
}

fn map_dispatch_error(e: &anyhow::Error) -> Result<Response<Body>, (StatusCode, String)> {
    if let Some(denied) = e.downcast_ref::<crate::services::gateway::service::PolicyDenied>() {
        return Err((StatusCode::FORBIDDEN, denied.to_string()));
    }
    if let Some(blocked) = e.downcast_ref::<crate::services::gateway::service::SafetyBlocked>() {
        return Err((StatusCode::FORBIDDEN, blocked.to_string()));
    }
    if let Some(quota) = e.downcast_ref::<crate::services::gateway::service::QuotaExceeded>() {
        let mut resp = build_error_response(StatusCode::TOO_MANY_REQUESTS, &quota.message);
        if let Ok(v) = HeaderValue::from_str(&quota.retry_after_seconds.to_string()) {
            resp.headers_mut().insert("retry-after", v);
        }
        return Ok(resp);
    }
    if let Some(upstream) = e.downcast_ref::<UpstreamError>() {
        return Err(map_upstream_error(upstream));
    }
    Err((StatusCode::BAD_GATEWAY, e.to_string()))
}

/// Translates an upstream failure into the status the *client* should see.
///
/// A malformed request (400/404/422) or rate limit (429) is the caller's to
/// fix, so it passes through; auth, 5xx, timeout, and transport failures are
/// the gateway/provider's problem and collapse to 502/504 without leaking the
/// upstream's credential-level detail.
pub fn map_upstream_error(e: &UpstreamError) -> (StatusCode, String) {
    let UpstreamError::Status {
        provider,
        status,
        message,
    } = e
    else {
        return (
            StatusCode::BAD_GATEWAY,
            "upstream provider unreachable".to_owned(),
        );
    };
    let mapped = match *status {
        400 | 404 | 422 => StatusCode::from_u16(*status).unwrap_or(StatusCode::BAD_REQUEST),
        429 => StatusCode::TOO_MANY_REQUESTS,
        408 | 504 => StatusCode::GATEWAY_TIMEOUT,
        _ => StatusCode::BAD_GATEWAY,
    };
    if mapped.is_server_error() {
        (mapped, "upstream provider error".to_owned())
    } else {
        (
            mapped,
            format!("{provider} rejected the request: {message}"),
        )
    }
}

pub(super) fn build_error_response(status: StatusCode, message: &str) -> Response<Body> {
    let escaped = message.replace('\\', "\\\\").replace('"', "\\\"");
    let body = format!(
        "{{\"type\":\"error\",\"error\":{{\"type\":\"api_error\",\"message\":\"{escaped}\"}}}}"
    );
    match Response::builder()
        .status(status)
        .header("content-type", "application/json")
        .body(Body::from(body))
    {
        Ok(resp) => resp,
        Err(e) => {
            tracing::error!(error = %e, status = %status, "Failed to build gateway error response");
            internal_error_response()
        },
    }
}

fn internal_error_response() -> Response<Body> {
    Response::builder()
        .status(StatusCode::INTERNAL_SERVER_ERROR)
        .header("content-type", "application/json")
        .body(Body::from(r#"{"error":"internal"}"#))
        .unwrap_or_else(|_| {
            let mut fallback = Response::new(Body::from(r#"{"error":"internal"}"#));
            *fallback.status_mut() = StatusCode::INTERNAL_SERVER_ERROR;
            fallback
        })
}