Skip to main content

Console tour

The EcoLink console lives at console.ecohash.com. This page maps every feature to where it lives.

The sidebar is grouped by task. Top-level items:

Overview​

  • Dashboard — landing page after login

Inference API​

  • Playground — interactive UI for every platform model (chat, vision, image, audio, video, embeddings, reranker, all in one page with tabs)
  • API Keys — create and revoke API keys for your account

Compute​

  • GPU Instances — your list of single-pod GPU dev environments
  • GPU Clusters — multi-replica container deployments
  • Model Instances — your own inference endpoints (user inference instances)

Storage​

  • Cloud Drives — block storage (attaches to one instance at a time)
  • Shared File Systems — network storage (multiple instances can attach simultaneously)

Registry​

  • Images — container images you've saved for reuse when launching instances
  • Models — registered models (HuggingFace-backed or filesystem-backed) you can deploy as inference instances

Account​

  • Users — team members and invites
  • Billing — credit balance, transactions, API usage
  • Messages — notifications (billing warnings, instance termination events, invite activity)

Top bar​

  • Account switcher (top-left) — if you belong to multiple accounts, switch between them here.
  • Messages bell (top-right) — shows unread notifications with a red dot + count. Click to see the 5 most recent; "View all" links to the Messages page.
  • User menu (top-right) — your email, log out.

Dashboard​

The first page after login — at-a-glance snapshot of everything running:

  • Current credit balance
  • Active GPU instances, clusters, inference instances
  • Recent activity / API usage
  • Notifications summary

Each card links to the full feature page.

Playground​

A single page with tabs for each model category:

  • Chat — text LLMs. Pick a model, type, get streamed responses. Includes token counts and latency metrics.
  • Vision — vision LLMs. Drag-and-drop an image + ask a question about it.
  • Image — text-to-image generation. Prompt, size selector, see the generated image.
  • Speech — text-to-speech and speech-to-text. Pick a voice, type text, play the audio; or upload a recording and see the transcript.
  • Video — text-to-video generation (async — takes 1–3 minutes).
  • Embeddings — paste text, see the vector.
  • Reranker — paste a query and candidate documents, see the reranked order.

Every call from the playground bills the same way as an API call. Most tabs include a Code panel showing the equivalent curl / Python / TypeScript / JavaScript you'd use to reproduce the request.

Compute pages​

  • GPU Instances — launch, list, extend, open terminal, open Jupyter, attach storage, view billing per instance.
  • GPU Clusters — multi-replica container deployments behind a single load-balanced URL.
  • Model Instances — inference endpoints you've launched. Each exposes an OpenAI-compatible URL with a model ID like <name>:<id>.

Storage pages​

  • Cloud Drives — create, list, delete block-storage drives. Attach to a GPU instance at launch.
  • Shared File Systems — create, list, delete network filesystems. Multiple instances can mount the same filesystem simultaneously.

Registry pages​

  • Images — container images. Save frequently-used images so you can pick from a dropdown at launch time instead of typing the full URL.
  • Models — user models. Register a HuggingFace repo or a folder on a shared filesystem; the entry is then usable when launching an inference instance.

Account pages​

  • Users — team members with roles (owner / member), pending invite links you can re-copy or revoke.
  • Billing — credit balance, full transaction history, per-model API usage breakdown.
  • Messages — every notification the platform has raised for your account.
  • Docs — you're reading them.
  • Support — ping the #ecolink-support Slack channel.

What to do next​