.png)
ElevenLabs Integration Guide
Connect enterprise workflows to ElevenLabs REST APIs for speech generation, transcription, dubbing, voice management, and audio processing.
ElevenLabs integration options at a glance
ElevenLabs primarily integrates through REST APIs for text-to-speech, speech-to-text, voice management, dubbing, audio isolation, sound effects, and conversational AI capabilities. Media-oriented endpoints can accept audio or video files, while selected long-running operations such as dubbing are handled asynchronously. Webhooks are available for selected products and events, but coverage is not universal. Martini can authenticate with the xi-api-key header, orchestrate API and file-processing workflows, map JSON and media metadata, expose controlled webhook or façade APIs, and schedule polling for operations without callbacks. Streaming designs should be validated against the selected endpoint and runtime requirements.
| Integration point | Supported by ElevenLabs? | Common use cases | How Martini supports it |
|---|---|---|---|
| REST APIs | Yes | Text-to-speech, speech-to-text, voice management, dubbing, audio isolation, sound effects, and conversational AI operations. | Martini can consume ElevenLabs REST endpoints, map request and response data, expose façade APIs, and orchestrate downstream processing. |
| Webhooks / outbound callbacks | Limited | Selected ElevenLabs products and events can notify an external endpoint when supported. | Martini can expose a controlled API endpoint or receive webhook requests through a workflow start trigger; event coverage and authenticity must be verified per product. |
| Bulk / async / batch APIs | Limited | Dubbing and other longer media operations can be submitted and tracked asynchronously; a universal bulk API was not confirmed. | Martini stores operation identifiers, polls status on a schedule, processes supported callbacks, and applies bounded retries and timeout handling. |
| File / attachment APIs | Yes | Speech-to-text, dubbing, and other media operations can accept audio or video inputs. | Martini can retrieve files, send multipart or binary requests, map media metadata, and route generated media or references to storage. |
| Streaming | Limited | Selected audio-generation and conversational use cases support HTTP streaming or WebSocket-based interactions. | Martini can support a streaming design only after the selected endpoint and runtime requirements are validated; standard HTTP APIs are the default approach. |
| Authentication | Yes | Standard API requests use an ElevenLabs API key, typically in the xi-api-key HTTP header. | Martini stores the key in secrets or protected environment configuration and injects it into outbound requests without exposing it in logs. |
| GraphQL APIs | Not confirmed | No official general-purpose ElevenLabs GraphQL API was confirmed. | Martini should use the documented REST APIs rather than assume GraphQL availability. |
| SOAP APIs | Not confirmed | No official ElevenLabs SOAP API was confirmed. | Martini should use REST, file, webhook, or scheduled API patterns where documented. |
How ElevenLabs exposes data and business events
ElevenLabs REST APIs
ElevenLabs documents REST endpoints for speech generation, transcription, voice operations, dubbing, audio processing, and other capabilities. Requests generally use an API key in the xi-api-key header and return JSON, media, or media references depending on the operation.
Martini implementation pattern
Martini workflows receive an application request or schedule, validate and transform the input, call the selected ElevenLabs endpoint, and route the response to storage or a business application. A Martini API can hide vendor-specific authentication and payload details from callers.
Implementation sequence
ElevenLabs file processing
Speech-to-text, dubbing, and other media-oriented operations may accept uploaded audio or video files. Supported formats, file sizes, and processing durations depend on the specific endpoint.
Martini implementation pattern
Martini retrieves or receives the source file, constructs the required multipart or binary request, and maps the returned transcript, project, audio, video, or metadata. Binary content is kept out of ordinary logs and may be routed to approved storage.
Implementation sequence
ElevenLabs asynchronous operations
Dubbing and other longer-running media operations may be accepted before processing is complete. A returned project or operation identifier is used to track progress.
Martini implementation pattern
Martini creates the operation, stores its identifier, and either processes a supported ElevenLabs event or runs a scheduled polling workflow. Completion, failure, timeout, and duplicate handling are modeled explicitly.
Implementation sequence
ElevenLabs selected webhooks
ElevenLabs documents webhook support for selected products and events rather than a universal event stream. Event types, schemas, authenticity, retry behavior, and delivery guarantees must be confirmed for the chosen feature.
Martini implementation pattern
Martini exposes a controlled REST endpoint or receives the webhook through a workflow start trigger, validates the request according to the applicable ElevenLabs guidance, deduplicates it, and starts downstream processing.
Implementation sequence
ElevenLabs streaming
Selected audio-generation and conversational use cases support HTTP streaming or WebSocket-based interactions. The exact interaction model varies by endpoint.
Martini implementation pattern
Martini can be considered for streaming integration only after endpoint behavior and runtime compatibility are verified. Where streaming is not suitable, a standard REST request or asynchronous workflow provides a more controlled design.
Implementation sequence
Common ElevenLabs integration patterns
Pattern 1: Transcribe customer-service recordings
When to use this pattern
Use this pattern when call recordings or other approved audio become available and transcripts must be attached to a customer, case, or ticket context. Validation should confirm the file and source correlation before downstream updates.
Integration direction
Example Mapping
| ElevenLabs Field | Canonical Field | Target Field |
|---|---|---|
| audio_file | sourceMediaReference | recordingReference |
| transcript_text | transcriptText | description |
| language_code | detectedLanguage | language |
| source_call_id | sourceCorrelationId | externalReference |
Martini implementation pattern
Martini retrieves the authorized file, submits it to speech-to-text, maps transcript metadata, validates required fields, and updates the target case or ticket. File errors, rate limits, and transient provider failures use retry or review branches, while the source correlation key prevents duplicate updates.
Martini capabilities used
- workflows
- API consumption
- file handling
- data mapping
- validation
- error handling
Pattern 2: Generate approved multilingual audio content
When to use this pattern
Use this pattern when approved text from a content or commerce application must become branded, localized, or accessible audio. Business rules can select a voice, model, language, and output format.
Integration direction
Example Mapping
| ElevenLabs Field | Canonical Field | Target Field |
|---|---|---|
| body | sourceText | text |
| locale | language | languageCode |
| voice_id | selectedVoiceId | voiceId |
| audio_reference | generatedMediaReference | assetReference |
Martini implementation pattern
Martini receives approved content, applies language and voice selection rules, calls text-to-speech, stores the generated result or reference, and updates the source system. Content hashes or source identifiers provide duplicate protection when a generation request is retried.
Martini capabilities used
- workflows
- API consumption
- business rules
- data mapping
- file and media routing
- retry handling
Pattern 3: Track asynchronous dubbing projects
When to use this pattern
Use this pattern for media localization where dubbing is accepted as a long-running project rather than completed in the initial request. It is appropriate when status, timeout, and failure states must be visible to an originating system.
Integration direction
Example Mapping
| ElevenLabs Field | Canonical Field | Target Field |
|---|---|---|
| source_media | sourceMediaReference | inputMedia |
| target_languages | requestedLanguages | languages |
| project_id | vendorOperationId | dubbingProjectId |
| status | processingStatus | assetStatus |
Martini implementation pattern
A Martini workflow creates the dubbing project, persists its identifier, and uses a supported event or scheduled polling to track status. On completion it retrieves the result and updates the originating system; bounded retries, duplicate-event protection, and timeout routing handle operational failures.
Martini capabilities used
- workflows
- API consumption
- scheduler triggers
- state tracking
- business rules
- error handling
Pattern 4: Expose an ElevenLabs audio gateway
When to use this pattern
Use this pattern when multiple internal or partner applications need a consistent API for speech generation or transcription without receiving ElevenLabs-specific credentials and payload details.
Integration direction
Example Mapping
| ElevenLabs Field | Canonical Field | Target Field |
|---|---|---|
| text | inputText | text |
| voice | requestedVoice | voice_id |
| output_format | outputFormat | output_format |
| result_reference | mediaReference | normalizedResponse |
Martini implementation pattern
Martini exposes a controlled REST API, authenticates callers using the organization’s security model, validates text and usage parameters, invokes ElevenLabs with a protected key, and returns a normalized response. Usage limits, correlation identifiers, sensitive-content controls, and provider-error mapping are applied centrally.
Martini capabilities used
- API exposure
- API consumption
- authentication and authorization
- validation
- data mapping
- business rules
- error handling
Applications commonly integrated with ElevenLabs
ElevenLabs can be incorporated into enterprise application landscapes where audio generation, transcription, dubbing, or voice-enabled workflows are required. These are standards-based integration patterns rather than claims of prebuilt ElevenLabs integrations.
| Application | Scenario | Direction | Martini Pattern |
|---|---|---|---|
| Salesforce | Transcribe call recordings, generate voice content for customer communications, or attach generated audio to customer-related records. | Salesforce → Martini → ElevenLabs | Martini receives a Salesforce request or recording reference, retrieves media when required, calls the appropriate ElevenLabs REST API, validates the result, and writes transcript or audio metadata back to Salesforce. |
| ServiceNow | Convert support-call recordings into transcripts and enrich incidents, cases, or related operational workflows. | ServiceNow → Martini → ElevenLabs | A Martini workflow retrieves approved recordings or references, submits them to speech-to-text, applies validation and routing rules, and updates ServiceNow with transcript metadata and processing status. |
| Zendesk | Associate voice-support transcripts with tickets and generate localized spoken support content. | Zendesk → Martini → ElevenLabs | Martini receives a ticket or recording event, submits supported media to ElevenLabs, normalizes the response, and updates the Zendesk ticket while routing failed jobs for retry or review. |
| Shopify | Generate product audio descriptions, accessibility content, or localized spoken content for product catalogs. | Shopify → Martini → ElevenLabs | Martini reads approved product content, selects a voice using language and brand rules, invokes text-to-speech, and returns generated-media metadata or a storage reference to Shopify. |
| Contentful | Turn approved editorial content into audio assets or localized media variants. | Contentful → Martini → ElevenLabs | A workflow watches for approved content, maps the text and voice parameters to ElevenLabs, stores the resulting media reference, and updates the Contentful publishing model after validation. |
| Moodle | Produce narrated course material, language variants, or transcripts for learning content. | Moodle → Martini → ElevenLabs | Martini retrieves approved learning content, applies language and voice rules, generates audio or transcripts, and publishes the resulting reference back to Moodle with status tracking. |
| Microsoft Teams | Process meeting or call recordings for transcription and downstream collaboration workflows. | Microsoft Teams → Martini → ElevenLabs | Martini consumes an authorized recording reference, submits the media to ElevenLabs, maps transcript output and correlation identifiers, and routes the result to collaboration or records workflows. |
| NetSuite | Generate audio-based product or operational content and associate resulting assets with items or transactions where useful. | NetSuite → Martini → ElevenLabs | Martini reads eligible NetSuite content, invokes ElevenLabs under configurable business rules, and writes output metadata to NetSuite or approved external media storage. |
How to build a ElevenLabs integration in Martini
Objective
Configure the ElevenLabs base URL and API key without embedding credentials in workflow definitions or exposing them in logs.
Instructions in Martini
- Store the xi-api-key value in Martini secrets or protected environment configuration
- Configure the REST request with the ElevenLabs authentication header
- Define separate configuration for development, testing, and production
Objective
Select an event, inbound API request, or schedule based on whether the use case is real-time, callback-driven, or asynchronous.
Instructions in Martini
- Use an API or application event for generation and transcription requests
- Expose a controlled webhook endpoint for supported ElevenLabs events
- Use a scheduler for operations without documented callbacks
Objective
Acquire text, audio, video, or vendor status data and preserve the source correlation key needed for downstream processing.
Instructions in Martini
- Retrieve authorized media before file submission
- Validate file type, size, duration, and required metadata
- Persist project or operation identifiers for asynchronous processing
Objective
Coordinate ElevenLabs calls, status checks, storage, target updates, and failure routes as a maintainable Martini workflow.
Instructions in Martini
- Call the appropriate REST endpoint
- Branch between immediate and asynchronous processing
- Apply timeout, retry, and failure handling paths
Objective
Convert application data into ElevenLabs request fields and normalize returned JSON, transcripts, media references, and status values.
Instructions in Martini
- Map text, voice, model, language, and output-format parameters
- Normalize transcript and project metadata for target applications
- Keep binary media out of ordinary application logs
Objective
Enforce business controls around approved content, voices, languages, usage, privacy, and duplicate processing.
Instructions in Martini
- Validate required fields before sending requests
- Select voices and formats using explicit business rules
- Use source identifiers or content hashes to prevent duplicate generation
Common ElevenLabs data objects used in integrations
| Object | Typical Use | Common target systems | Martini handling |
|---|---|---|---|
| Voices | Voice selection, voice settings, language routing, and management of synthesis options. | Contentful, Shopify, Salesforce, Moodle, configuration stores | Martini retrieves and caches relevant voice identifiers, applies selection rules, and maps voice settings into generation requests. |
| Text-to-speech generations | Audio produced from text, voice, model, language, and output-format parameters. | Object storage, Shopify, Contentful, Moodle, Salesforce | Martini validates inputs, submits the request, tracks correlation data, and stores audio or a returned media reference without logging sensitive binary content. |
| Speech-to-text transcriptions | Transcript text, language, segments, and related metadata derived from audio. | Salesforce, ServiceNow, Zendesk, Microsoft Teams, data stores | Martini uploads supported media, normalizes transcript responses, applies validation, and updates downstream applications. |
| Dubbing projects | Asynchronous translation or recreation of audio and video in target languages. | Content management systems, media storage, publishing platforms | Martini creates the project, persists its identifier, polls or receives a supported event, and retrieves completed media with timeout and retry controls. |
| Projects | Longer-form text-to-speech content and generated audio assets where available for the account and product configuration. | Contentful, Moodle, media stores, internal content services | Martini treats project processing as potentially asynchronous, maps status and asset references, and routes failures for review. |
| Conversational AI agents | Configured agents used in voice-based conversational applications. | Internal applications, customer-service platforms, API gateways | Martini can expose or consume API operations around confirmed agent capabilities, protect credentials, and apply caller authorization and usage rules. |
Authentication and security considerations
API key protection
ElevenLabs standard API access uses an API key, commonly sent in the xi-api-key header. Store it in Martini secrets or protected environment configuration rather than workflow definitions.
Media and privacy controls
Recordings, transcripts, and generated audio may contain personal or regulated information. Apply retention, deletion, access, and logging policies appropriate to the content.
- Restrict API keys to the minimum available permissions and rotate them periodically.
- Do not expose credentials or sensitive media in workflow logs or public API responses.
- Validate webhook authenticity and reject unexpected or replayed requests where supported.
Operational considerations for ElevenLabs integrations
Reliability and scale
Confirm account, workspace, endpoint, and plan-specific quotas before controlling concurrency. Use exponential backoff for 429 responses and transient service errors, and queue high-volume generation or transcription work where appropriate.
Media and asynchronous processing
Validate formats, file sizes, and durations before upload. Persist dubbing project identifiers, poll with bounded retries when callbacks are unavailable, and route failed or timed-out jobs for review.
Data integrity
- Preserve pagination cursors for voices, projects, history, and other collections.
- Use source identifiers, content hashes, or operation IDs to provide idempotency where the vendor endpoint does not.
- Keep API paths, versions, voice identifiers, models, languages, and output formats configurable.
- Test response mappings when ElevenLabs schemas or product capabilities change.
Why use Martini instead of scripts or point-to-point integrations?
Orchestration instead of isolated scripts
Martini centralizes ElevenLabs authentication, API calls, file handling, transformations, business rules, asynchronous status tracking, and downstream updates in maintainable workflows.
Reusable enterprise controls
Teams can expose a consistent API façade, protect vendor credentials, standardize correlation and duplicate handling, and route errors without replicating logic across every consuming application.
- Combine API, webhook, schedule, storage, and application steps in one integration flow.
- Separate vendor-specific payloads from internal canonical models.
- Apply validation, retries, monitoring, and deployment configuration consistently.
Frequently asked questions
ElevenLabs integrates primarily through REST APIs for text-to-speech, speech-to-text, voice management, dubbing, audio processing, and related capabilities. Media endpoints can accept audio or video files, selected products support webhooks, and asynchronous operations can be tracked through status polling or supported callbacks.
Yes. Martini can consume ElevenLabs REST APIs, authenticate with the xi-api-key header, submit supported media files, transform JSON and media metadata, expose façade APIs, and receive selected ElevenLabs webhook events. A dedicated native Martini connector is not documented in the supplied sources.
No. A dedicated ElevenLabs connector is not required. Martini can use ElevenLabs’ confirmed native REST APIs, file-processing endpoints, API-key authentication, selected webhooks, and scheduled polling patterns.
Lonti does not charge an additional per-connector or per-vendor fee to integrate ElevenLabs. The integration is subject to the provisioned capacity of the Martini environment. Separate costs may apply from ElevenLabs, cloud infrastructure, storage providers, or other third parties based on subscription, usage, and deployment model.
Use the documented REST APIs as the primary mechanism. Use file uploads for supported audio and video operations, selected webhooks when the relevant product and event are covered, and scheduled polling for asynchronous operations without a callback. GraphQL and SOAP were not confirmed.
ElevenLabs documents webhook support for selected products and events, not every API request or operation. Coverage, event schemas, authenticity requirements, retries, and delivery guarantees must be verified for the specific feature; Martini can receive supported callbacks or poll status instead.
Martini can initiate real-time API workflows, process supported webhook notifications, or schedule polling for asynchronous projects. It maps voices, generations, transcriptions, dubbing projects, statuses, and media references into target applications while retaining correlation identifiers and applying duplicate protection.
Martini can validate inputs, route authentication and media errors, retry transient failures with bounded backoff, and handle rate limits and provider errors through workflow branches. Source identifiers, content hashes, operation IDs, and stored status can prevent duplicate generation or repeated downstream updates.
Yes. Martini can expose a controlled REST API that validates caller input, applies organization-specific authentication and business rules, invokes ElevenLabs with a protected API key, normalizes the response, and hides vendor-specific credentials and payload details from consuming applications.
Related Martini documentation
Workflows
Connect ElevenLabs to your enterprise workflows
Use Martini to integrate ElevenLabs REST APIs, media processing, selected webhooks, and asynchronous operations with the applications and data stores your organization already uses.