Integration

DeutschlandGPT integration

DeutschlandGPT Platform API, an OpenAI-compatible AI API hosted in Germany: chat completions and stateless responses with GPT, Claude, Gemini, Mistral and open models, embeddings, live web search, image generation and editing, text-to-video and avatar videos, text-to-speech, music and transcription, plus document folders with semantic search, Custom GPTs and workspace usage reports. Every call is billed to the workspace's prepaid EUR credits (per token, image, second or character, web search per request); API access is in every plan and needs the role permission to use the Platform API. Not exposed: streaming, the Anthropic-format endpoints, API-key listing, voice cloning and photo avatar creation. Authentication is an API key (sk_...) sent as a Bearer token.

Connect DeutschlandGPT How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in DeutschlandGPT.Built from the 41 DeutschlandGPT tools below, on the connection you add once.

What you can do with DeutschlandGPT

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Document folder, 6 operations

create_document_folderCreate document folderPOSTCreates a document folder in your workspace; add files with upload_document_folder_file and search them with search_document_folders.
delete_document_folderDelete document folderDELETEPermanently deletes a document folder and every file in it.
get_document_folderGet document folderGETReturns one document folder with its files (id, name, type, creation date).
list_document_foldersList document foldersGETLists your workspace's document folders with their file counts. Without query every folder is returned; with query, folders are filtered by name and limit and offset page the results.
search_document_foldersSearch document foldersPOSTSemantic search over the files in your document folders: returns the most relevant text chunks with file name, folder id and similarity score. Respects the key's folder permissions.
update_document_folderUpdate document folderPUTRenames a document folder or changes its description; leave out a field to keep its current value.

Custom gpt, 5 operations

create_custom_gptCreate custom gptPOSTCreates a Custom GPT (an assistant with its own instructions, default model and files) in your workspace; workspace members can use it.
delete_custom_gptDelete custom gptDELETEPermanently deletes a Custom GPT.
get_custom_gptGet custom gptGETReturns one Custom GPT with its instructions, settings and files.
list_custom_gptsList custom gptsGETLists your workspace's Custom GPTs (assistants with their own instructions, files and model); pass an id as custom_gpt_id to create_chat_completion to answer with one.
update_custom_gptUpdate custom gptPUTChanges a Custom GPT's configuration; only the fields you send are updated, leave out a field to keep its current value.

Usage, 4 operations

get_daily_usageGet daily usageGETReturns per-day usage (requests, tokens, text, image and embedding cost in micro-cents) over a date range (default last 30 days). Needs an admin API key.
get_usage_by_api_keyGet usage by api keyGETReturns usage per API key (requests, cost in micro-cents) over a date range (default last 30 days); key names come back as SHA-256 hashes. Needs an admin API key.
get_usage_by_modelGet usage by modelGETReturns usage per model (requests, tokens, cost in micro-cents) over a date range (default last 30 days). Needs an admin API key.
get_usage_summaryGet usage summaryGETReturns total cost, requests, tokens and active API keys for the workspace over a date range (default last 30 days). Needs a key with the admin:export_usage scope. Costs are in micro-cents.

Document file, 3 operations

delete_document_folder_fileDelete document folder fileDELETEPermanently deletes one file from a document folder.
replace_document_folder_fileReplace document folder filePUTReplaces a file in a document folder with a new upload (max 50 MB); if the upload fails or is refused, the original file stays. A file that is not in this folder returns 404.
upload_document_folder_fileUpload document folder filePOSTUploads a file (max 50 MB) to a document folder; returns file_id. Indexing for search_document_folders runs asynchronously, so a new file may take a moment to become searchable.

Model, 3 operations

get_modelGet modelGETReturns one text model by id (owner and registration time).
list_all_modelsList all modelsGETLists every model of every kind (text, image, video, avatar video, speech, transcription, embedding) with its kind, provider, host region and, for speech models, voices. Pass an id as model to the matching tool.
list_modelsList modelsGETLists the text models your key can use; pass an id as model to create_chat_completion or create_response. For image, video, audio and embedding models use list_all_models.

Avatar, 2 operations

delete_avatarDelete avatarDELETEPermanently deletes a custom avatar from your workspace (it is removed from the provider too). Takes the registry id from list_avatars (id, not avatar_id).
list_avatarsList avatarsGETLists the custom avatars owned by your workspace; pass an avatar_id to create_avatar_video. The models' built-in avatars are not listed.

Avatar video, 2 operations

create_avatar_videoCreate avatar videoPOSTRenders a talking-head avatar video that reads your script. Credits are pre-charged from the script length and reconciled to the real duration. Waits for the job and returns the finished result; gives up after 10 minutes.
get_avatar_video_jobGet avatar video jobGETReturns an avatar video job's status (pending, succeeded or failed); when succeeded, data.url holds a signed video link valid for expires_in_seconds. create_avatar_video already waits; use this to get a fresh link.

Custom gpt file, 2 operations

delete_custom_gpt_fileDelete custom gpt fileDELETEDetaches a file from a Custom GPT; the file itself is not deleted.
upload_custom_gpt_fileUpload custom gpt filePOSTUploads a file (max 50 MB) and attaches it to a Custom GPT; returns file_id. A Custom GPT holds at most 20 files, at most 10 of them images.

Image, 2 operations

create_imageCreate imagePOSTGenerates images from a text prompt and returns them as base64 (data[].b64_json). Billed per image from prepaid credits, by model, quality and size. Leave out model for dall-e-3; list image models with list_all_models.
edit_imageEdit imagePOSTEdits an uploaded image from a prompt (optionally with a PNG mask) and returns base64 images. Billed per image from prepaid credits. Use a model whose supports_edits is true in list_all_models.

Video, 2 operations

create_videoCreate videoPOSTGenerates videos from a text prompt (default model veo-3.1-fast). Credits are charged up front by resolution, duration and count; failures are refunded. Waits for the job and returns the finished result; gives up after 10 minutes.
get_video_jobGet video jobGETReturns a video job's status (pending, succeeded or failed); when succeeded, data[].url holds signed video links valid for expires_in_seconds. create_video already waits; use this to get fresh links.

Voice, 2 operations

delete_voiceDelete voiceDELETEPermanently deletes a cloned voice from your workspace (it is removed from ElevenLabs too). Takes the registry id from list_voices (id, not voice_id).
list_voicesList voicesGETLists the cloned voices owned by your workspace; pass a voice_id as voice to create_speech or as voice_id to create_avatar_video. Public ElevenLabs voices are not listed.

Audit log, 1 operation

get_audit_logsGet audit logsGETReturns the enterprise workspace's audit log, newest first, filterable by category, action, actor and dates; actor emails and target labels come back as SHA-256 hashes. Needs an admin key.

Chat completion, 1 operation

create_chat_completionCreate chat completionPOSTGenerates a chat reply with any text model (model id from list_models). Billed per input and output token from prepaid credits. Pass custom_gpt_id to answer with a Custom GPT. No streaming; use response_format for JSON output.

Embedding, 1 operation

create_embeddingCreate embeddingPOSTTurns one or more texts into embedding vectors, one per input, for semantic search or clustering. Billed per input token from prepaid credits. Default model text-embedding-3-large.

Music, 1 operation

create_musicCreate musicPOSTGenerates a music track from a text prompt and returns the audio file (mp3 by default). Billed per minute of requested duration from prepaid credits; leave out duration_ms for 30 seconds.

Response, 1 operation

create_responseCreate responsePOSTGenerates a model reply with the stateless Responses API (model id from list_models). Billed per token from prepaid credits. Send the whole conversation in input every time; earlier responses are not stored.

Speech, 1 operation

create_speechCreate speechPOSTConverts text to speech and returns the audio file (mp3 by default). Billed per character from prepaid credits. Pick the model first, then one of ITS voices from list_all_models; leave out voice for the model default.

Transcription, 1 operation

transcribe_audioTranscribe audioPOSTTranscribes an audio file (max 25 MB) to text. Billed per audio second from prepaid credits. Use response_format verbose_json with diarize true for speaker turns; default model voxtral-mini-2507.

Web search, 1 operation

search_webSearch webPOSTRuns a live web search and returns ranked results with title, URL, snippet and age. Billed a flat amount per request from prepaid credits, whatever count is.

Use DeutschlandGPT from Claude Code, Cursor or Codex

Connect DeutschlandGPT in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its DeutschlandGPT operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect DeutschlandGPT

Create your workspace
Sign up free. Nothing to install.
Connect DeutschlandGPT once
Add your DeutschlandGPT connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.