Speech-to-text API · served now
Whisper transcription through the OpenAI-compatible /v1/audio/transcriptions endpoint, served from hardware AxForge owns in Sweden. Your audio is processed in memory and retained nowhere.
Specifications
| Model | Whisper — speech-to-text |
|---|---|
| API model name | whisper |
| Endpoint | POST /v1/audio/transcriptions |
| Request shape | Multipart upload — file + model, OpenAI-compatible |
| Hardware | NVIDIA DGX Spark (GB10, 128 GB unified memory) — owned and operated by AxForge |
| Region | eu-se-1 · Stockholm, Sweden — pinned with x-axforge-region, echoed on the response |
| Pricing | €0.005 / minute (launch pricing) |
For quality benchmarks, see the official model card — we publish only numbers we measured ourselves.
Quickstart
$ export OPENAI_BASE_URL=https://api.axforge.ai/v1 $ curl -sS "$OPENAI_BASE_URL/audio/transcriptions" \ -H "Authorization: Bearer $AXFORGE_API_KEY" \ -H "x-axforge-region: eu-se-1" \ -F model=whisper \ -F file=@meeting.mp3
# pip install openai from openai import OpenAI client = OpenAI( base_url="https://api.axforge.ai/v1", api_key="YOUR_AXFORGE_KEY", default_headers={"x-axforge-region": "eu-se-1"}, ) with open("meeting.mp3", "rb") as audio: r = client.audio.transcriptions.create(model="whisper", file=audio) print(r.text)
Data & privacy
Prompts and completions are processed in memory in Sweden — not written to disk, not logged, not retained, and never used to train anything. For this endpoint that means the audio you upload and the transcript we return. We keep only request metadata (token counts, timestamps, status) for billing and operations. The full policy is at axforge.ai/privacy.
FAQ
Yes. It is the standard /v1/audio/transcriptions shape — a
multipart upload with model=whisper — so the official OpenAI
SDKs work by changing base_url to
https://api.axforge.ai/v1.
In Stockholm, Sweden (region eu-se-1), on hardware AxForge
owns and operates. TLS terminates in the EU. Spain (eu-es-1) is
live for the platform; more EU regions are in deployment.
No. Requests are processed in memory under our zero-retention policy — the audio you send and the transcript we return are not written to disk, not logged, not retained, and never used for training. See the privacy policy.
Yes — that is exactly what it is built for. The service is GDPR compliant: EU region, EU-owned hardware, zero retention. The privacy policy describes the full data handling.
€0.005 per minute of audio (launch pricing). For committed volume, talk to an engineer.