🔒 feat: Add allowedAddresses Exemption to Speech (STT/TTS) and OCR Config Schemas (#14559)

* feat: add allowedAddresses exemption to speech (STT/TTS) and OCR config schemas

Add the existing allowedAddressesSchema as an optional field on sttSchema,
ttsSchema, and ocrSchema, reusing the schema already attached to endpoints,
mcpSettings, and actions so port scoping and normalization stay identical.

STT and TTS resolve a single provider by counting non-empty section keys, so
exclude the allowedAddresses key from that scan. Without the exclusion a
configured exemption list would be counted as a second provider and trip the
"Multiple providers are set" guard. The field is inert on its own: nothing
reads it for SSRF yet, and provider detection now ignores it.

* fix: preserve allowedAddresses through the OCR config loaders

loadOCRConfig rebuilt the ocr config with only apiKey, baseURL,
mistralModel, and strategy, dropping allowedAddresses before it reached
req.config.ocr. Pass it through in both the AppService loader
(packages/data-schemas/src/app/ocr.ts) and the duplicate at
packages/api/src/files/ocr.ts so the exemption survives config load.
This commit is contained in:
Dustin Healy 2026-08-05 10:42:43 -07:00 committed by GitHub
parent 3f0a1ec8d9
commit 4cec1a675f
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
9 changed files with 159 additions and 4 deletions

View file

@ -157,7 +157,7 @@ class STTService {
}
const providers = Object.entries(sttSchema).filter(
([, value]) => Object.keys(value).length > 0,
([key, value]) => key !== 'allowedAddresses' && Object.keys(value).length > 0,
);
if (providers.length !== 1) {

View file

@ -10,7 +10,7 @@ jest.mock('librechat-data-provider', () => ({
}));
jest.mock('~/server/services/Config', () => ({ getAppConfig: jest.fn() }));
const { getFileExtensionFromMime, MIME_TO_EXTENSION_MAP } = require('./STTService');
const { STTService, getFileExtensionFromMime, MIME_TO_EXTENSION_MAP } = require('./STTService');
describe('getFileExtensionFromMime', () => {
it('should normalize audio/x-m4a to m4a', () => {
@ -51,6 +51,41 @@ describe('getFileExtensionFromMime', () => {
});
});
describe('STTService.getProviderSchema provider detection', () => {
const service = new STTService();
const buildReq = (stt) => ({ config: { speech: { stt } } });
it('resolves exactly one provider when allowedAddresses is set alongside it', async () => {
const req = buildReq({
allowedAddresses: ['127.0.0.1:8080'],
openai: { url: 'http://127.0.0.1:8080', apiKey: 'sk', model: 'whisper-1' },
});
const [provider, schema] = await service.getProviderSchema(req);
expect(provider).toBe('openai');
expect(schema.url).toBe('http://127.0.0.1:8080');
});
it('reports "No provider is set" when only allowedAddresses is present', async () => {
const req = buildReq({ allowedAddresses: ['127.0.0.1:8080'] });
await expect(service.getProviderSchema(req)).rejects.toThrow('No provider is set');
});
it('reports "Multiple providers" when two providers are set even with allowedAddresses', async () => {
const req = buildReq({
allowedAddresses: ['127.0.0.1:8080'],
openai: { url: 'http://127.0.0.1:8080', apiKey: 'sk', model: 'whisper-1' },
azureOpenAI: {
instanceName: 'inst',
apiKey: 'sk',
deploymentName: 'dep',
apiVersion: '2024',
},
});
await expect(service.getProviderSchema(req)).rejects.toThrow('Multiple providers are set');
});
});
describe('STT audio format validation with MIME normalization', () => {
const acceptedFormats = ['flac', 'mp3', 'mp4', 'mpeg', 'mpga', 'm4a', 'ogg', 'wav', 'webm'];

View file

@ -52,7 +52,7 @@ class TTSService {
);
}
const providers = Object.entries(ttsSchema).filter(
([, value]) => Object.keys(value).length > 0,
([key, value]) => key !== 'allowedAddresses' && Object.keys(value).length > 0,
);
if (providers.length !== 1) {

View file

@ -23,7 +23,7 @@ jest.mock('./streamAudio', () => ({
jest.mock('~/server/services/Config', () => ({ getAppConfig: jest.fn() }));
const { resolveConfigSecret } = require('@librechat/api');
const { TTSService } = require('./TTSService');
const { TTSService, getProvider } = require('./TTSService');
describe('TTSService provider header construction with an undecryptable apiKey', () => {
let service;
@ -70,3 +70,35 @@ describe('TTSService provider header construction with an undecryptable apiKey',
expect(headers).not.toHaveProperty('Authorization');
});
});
describe('TTSService getProvider detection', () => {
const buildConfig = (tts) => ({ speech: { tts } });
it('resolves exactly one provider when allowedAddresses is set alongside it', async () => {
const provider = await getProvider(
buildConfig({
allowedAddresses: ['localhost:11434'],
localai: { url: 'http://localhost:11434/tts', apiKey: 'sk' },
}),
);
expect(provider).toBe('localai');
});
it('reports "No provider is set" when only allowedAddresses is present', async () => {
await expect(
getProvider(buildConfig({ allowedAddresses: ['localhost:11434'] })),
).rejects.toThrow('No provider is set');
});
it('reports "Multiple providers" when two providers are set even with allowedAddresses', async () => {
await expect(
getProvider(
buildConfig({
allowedAddresses: ['localhost:11434'],
openai: { url: 'http://localhost:11434', apiKey: 'sk' },
localai: { url: 'http://localhost:11434/tts', apiKey: 'sk' },
}),
),
).rejects.toThrow('Multiple providers are set');
});
});

View file

@ -11,5 +11,6 @@ export function loadOCRConfig(config?: TCustomConfig['ocr']): TCustomConfig['ocr
baseURL,
mistralModel,
strategy: config?.strategy ?? OCRStrategy.MISTRAL_OCR,
allowedAddresses: config?.allowedAddresses,
};
}

View file

@ -492,6 +492,62 @@ describe('allowedAddressesSchema', () => {
});
expect(result.success).toBe(false);
});
it('accepts the field on speech.stt', () => {
const result = configSchema.safeParse({
version: '1.0',
speech: { stt: { allowedAddresses: ['127.0.0.1:8080'] } },
});
expect(result.success).toBe(true);
});
it('accepts the field on speech.tts', () => {
const result = configSchema.safeParse({
version: '1.0',
speech: { tts: { allowedAddresses: ['localhost:11434', 'ollama.internal:11434'] } },
});
expect(result.success).toBe(true);
});
it('accepts the field on ocr', () => {
const result = configSchema.safeParse({
version: '1.0',
ocr: { allowedAddresses: ['10.0.0.5:443'] },
});
expect(result.success).toBe(true);
});
it('omitting the field on ocr leaves it undefined', () => {
const result = configSchema.safeParse({ version: '1.0', ocr: {} });
expect(result.success).toBe(true);
if (result.success) {
expect(result.data.ocr?.allowedAddresses).toBeUndefined();
}
});
it('rejects a public IP at the speech.stt location', () => {
const result = configSchema.safeParse({
version: '1.0',
speech: { stt: { allowedAddresses: ['8.8.8.8:53'] } },
});
expect(result.success).toBe(false);
});
it('rejects a bare host at the speech.tts location', () => {
const result = configSchema.safeParse({
version: '1.0',
speech: { tts: { allowedAddresses: ['localhost'] } },
});
expect(result.success).toBe(false);
});
it('rejects a CIDR range at the ocr location', () => {
const result = configSchema.safeParse({
version: '1.0',
ocr: { allowedAddresses: ['10.0.0.0/24'] },
});
expect(result.success).toBe(false);
});
});
});

View file

@ -1200,6 +1200,7 @@ const ttsLocalaiSchema = z.object({
});
const ttsSchema = z.object({
allowedAddresses: allowedAddressesSchema,
openai: ttsOpenaiSchema.optional(),
azureOpenAI: ttsAzureOpenAISchema.optional(),
elevenlabs: ttsElevenLabsSchema.optional(),
@ -1222,6 +1223,7 @@ const sttAzureOpenAISchema = z.object({
});
const sttSchema = z.object({
allowedAddresses: allowedAddressesSchema,
openai: sttOpenaiSchema.optional(),
azureOpenAI: sttAzureOpenAISchema.optional(),
});
@ -1788,6 +1790,7 @@ export const webSearchSchema = z.object({
export type TWebSearchConfig = DeepPartial<z.infer<typeof webSearchSchema>>;
export const ocrSchema = z.object({
allowedAddresses: allowedAddressesSchema,
mistralModel: z.string().optional(),
apiKey: z.string().optional().default('${OCR_API_KEY}'),
apiKeyPreview: apiKeyPreviewSchema,

View file

@ -0,0 +1,27 @@
import { OCRStrategy } from 'librechat-data-provider';
import { loadOCRConfig } from './ocr';
describe('loadOCRConfig', () => {
it('returns undefined when no config is provided', () => {
expect(loadOCRConfig(undefined)).toBeUndefined();
});
it('preserves allowedAddresses so the exemption survives config load', () => {
const loaded = loadOCRConfig({
apiKey: '${OCR_API_KEY}',
baseURL: 'https://ocr.internal:8080',
strategy: OCRStrategy.MISTRAL_OCR,
allowedAddresses: ['ocr.internal:8080'],
});
expect(loaded?.allowedAddresses).toEqual(['ocr.internal:8080']);
});
it('leaves allowedAddresses undefined when it is not configured', () => {
const loaded = loadOCRConfig({
apiKey: 'key',
baseURL: 'https://api.mistral.ai',
strategy: OCRStrategy.MISTRAL_OCR,
});
expect(loaded?.allowedAddresses).toBeUndefined();
});
});

View file

@ -11,5 +11,6 @@ export function loadOCRConfig(config?: TCustomConfig['ocr']): TCustomConfig['ocr
baseURL,
mistralModel,
strategy: config?.strategy ?? OCRStrategy.MISTRAL_OCR,
allowedAddresses: config?.allowedAddresses,
};
}