OceanToken

ai.oceantokenPublished by oceantoken.aiOfficial publisherv1.0.1Updated Oct 24, 2026

500+ AI models in one account: find and price models, make images, video, speech and text.

VerifiedStreamable HTTPWeb executableMedia & DesignAI & ML

Overview

AI-generated overview

OceanToken gives an assistant one account for 500+ AI models: search and price models, generate images, video, speech and text, and transcribe audio.

What you can do
  • Search the model catalog and check pricing, limits and accepted inputs
  • Estimate a request's cost and whether the account balance covers it
  • Generate images, create video jobs and poll their status, produce speech
  • Transcribe audio into transcripts and SRT/VTT subtitles
  • Delegate sub-tasks to other models and upload local files as references
  • Check account balance, available balance and spend
When to use it
Useful when a coding agent needs capabilities it lacks itself: image, video, voice-over or subtitle generation, access to models it does not have, or cost estimates before spending. Also handy for delegating a sub-task to a cheaper model or getting a second opinion.
Requirements
Remote streamable HTTP endpoint; no local package. An OceanToken account with prepaid credit is required. Authentication is OAuth with dynamic client registration, or an API key sent as a bearer token (for example via OCEANTOKEN_API_KEY). Network access to the OceanToken endpoint and API.
Before you install
Usage is billed from a prepaid balance, and video jobs place a hold when they start; failed jobs are not charged. Prompts and uploaded files are sent to the chosen model through OceanToken, and generated files are stored in OceanToken storage and shared via time-limited links. Sign-in creates an API key that can be revoked; an optional credit limit can cap agent spending.

Installation

In SourceWeft

  1. Open OceanToken in the dashboard and add it to a workspace.
  2. Connect the required credentials. They are stored privately and never shown on public pages.
  3. Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Use with an agent

Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.

Read https://sourceweft.com/mcp/ai.oceantoken%2Fmcp/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.

Read the install guide the agent follows

Other MCP clients

Add this to your client's mcpServers config. The client completes OAuth sign-in on first connect.

{
  "mcpServers": {
    "oceantoken": {
      "type": "http",
      "url": "https://mcp.oceantoken.ai/mcp"
    }
  }
}

README

[OceanToken]

OceanToken plugins for coding agents

OceanToken is one account for 500+ AI models: GPT, Claude, Gemini, DeepSeek and Qwen for text; GPT Image, Seedream, Recraft and Gemini for images; Seedance, Veo, Sora and Wan for video; plus text-to-speech and transcription. This plugin brings all of it into your coding agent: one command connects Codex, Claude Code, Cursor, Gemini CLI, OpenClaw, OpenCode, VS Code, WorkBuddy, Hermes Agent and Windsurf, and any other MCP client can add the server by hand. Your agent can find a model, check what it costs, generate the asset, and save it into your project.

What your agent gets

ToolWhat it does
search_modelsSearch the catalog by name, type, publisher, capability and price.
get_modelPricing, limits, and the sizes, durations and inputs a model accepts.
estimate_costWhat a request will cost before it runs, and whether your balance covers it.
generate_imageText-to-image and image edits, with previews and download links.
create_video + get_jobVideo from text, from a first (and last) frame, or from reference media.
generate_speechVoice-overs and narration.
transcribe_audioTranscripts and SRT/VTT subtitles.
ask_modelHand a sub-task to another model: a cheaper one, or a second opinion.
upload_fileUse local files as references.
get_accountBalance, available balance and spend.

Three skills teach the agent how to use them well:

  • oceantoken-media: produce images, video and audio for a project.
  • oceantoken-models: choose, price and delegate.
  • oceantoken-setup: run Codex or Claude Code itself on OceanToken, or wire up your app.

The same plugin folder carries manifests for each host (.codex-plugin, .claude-plugin, .cursor-plugin, and the open Agent Plugins plugin.json), so every host gets the same tools and skills.

Install

You need an OceanToken account (sign up) with some credit.

One command, any agent

Ask your agent:

Install OceanToken for me by following https://mcp.oceantoken.ai/install

or run it yourself:

bash
npx -y @oceantoken/cli@latest connect <client>

<client> is one of codex, claude-code, cursor, gemini, openclaw, opencode, vscode, workbuddy, hermes or windsurf. It installs the plugin or MCP entry and the three skills (Windsurf / Devin Desktop has no skills folder, so it gets the server only), and starts the sign-in. See cli/ for options (API keys, dry run, disconnect, and use to run Codex or Claude Code itself on OceanToken models) and the guide for a walkthrough. The sections below do the same by hand.

Codex

bash
codex plugin marketplace add OceanTokenAI/oceantoken-pluginscodex plugin add oceantoken@oceantokencodex mcp login oceantoken      # opens the OceanToken sign-in page

In the Codex app you can also add the marketplace, then install OceanToken from the Plugins panel and sign in when asked.

Claude Code

text
/plugin marketplace add OceanTokenAI/oceantoken-plugins/plugin install oceantoken@oceantoken/mcp                             # pick "plugin:oceantoken:oceantoken" and authenticate

Cursor

[Add OceanToken to Cursor]

The button adds the MCP server; Cursor asks you to sign in the first time a tool runs. To get the skills as well, install the plugin from this repository (it has a .cursor-plugin marketplace manifest).

Gemini CLI

bash
gemini extensions install https://github.com/OceanTokenAI/oceantoken-plugins

Then, inside Gemini CLI, run /mcp auth oceantoken and sign in. The extension brings the MCP server, the three skills and a short GEMINI.md.

OpenClaw

bash
openclaw plugins install clawhub:@oceantoken/oceantoken --accept-capabilitiesopenclaw mcp set oceantoken '{"url":"https://mcp.oceantoken.ai/mcp","transport":"streamable-http","auth":"oauth"}'openclaw mcp login oceantoken

The ClawHub package brings the three skills (also on ClawHub on their own: oceantoken-media, oceantoken-models, oceantoken-setup). The mcp set entry registers the server with OpenClaw so mcp login can sign you in; it replaces the plugin's own unauthenticated entry of the same name.

Any other MCP client

OceanToken is in the official MCP Registry as ai.oceantoken/mcp, so clients that browse the registry (VS Code and GitHub Copilot, for example) can find it by name.

The server is https://mcp.oceantoken.ai/mcp (streamable HTTP). It supports OAuth with dynamic client registration, so most clients just need the URL. To use an API key instead, send it as a bearer token:

bash
# Codex, without the pluginexport OCEANTOKEN_API_KEY=sk-...codex mcp add oceantoken --url https://mcp.oceantoken.ai/mcp --bearer-token-env-var OCEANTOKEN_API_KEY
# Claude Code, without the plugin. The single quotes keep the key off the command line:# Claude Code reads OCEANTOKEN_API_KEY from its environment when it starts.claude mcp add --transport http oceantoken https://mcp.oceantoken.ai/mcp \  --header 'Authorization: Bearer ${OCEANTOKEN_API_KEY}'
json
{  "mcpServers": {    "oceantoken": { "type": "http", "url": "https://mcp.oceantoken.ai/mcp" }  }}

Muse Code (Meta)

Add the server to the mcp_servers block of your Muse Code settings, then sign in:

json
{  "mcp_servers": {    "oceantoken": { "transport": "streamable_http", "url": "https://mcp.oceantoken.ai/mcp" }  }}
bash
muse mcp login oceantoken

Muse Code also reads skills from ~/.agents/skills: copy the folders in plugins/oceantoken/skills/ there to get the OceanToken workflows.

Make your agent reach for OceanToken

Agents choose their own tools. They use OceanToken for anything they cannot do themselves (video, voice-overs, subtitles, models they don't have, cost estimates). For images, Codex also has a built-in generator, so ask for OceanToken by name, or make it a rule:

  • Ask explicitly: $oceantoken-media ... in Codex, /oceantoken:oceantoken-media ... in Claude Code, or simply "use OceanToken to ...".

  • Make it a project rule in AGENTS.md (Codex) or CLAUDE.md (Claude Code):

    markdown
    Generate images, video, voice-overs and subtitles with the OceanToken tools.Estimate the cost first and ask me before spending more than $2.

  • Skip approval prompts for the read-only tools in Codex (~/.codex/config.toml):

    toml
    [plugins."oceantoken@oceantoken".mcp_servers.oceantoken.tools.search_models]approval_mode = "approve"[plugins."oceantoken@oceantoken".mcp_servers.oceantoken.tools.estimate_cost]approval_mode = "approve"[plugins."oceantoken@oceantoken".mcp_servers.oceantoken.tools.get_job]approval_mode = "approve"

Signing in

The sign-in page offers two ways to connect:

  • Sign in: your OceanToken email and password. This creates an API key for the connecting app, which you can see and revoke under API keys in the console.
  • API key: paste a key you already have. Give it a credit limit if you want a cap for the agent.

Revoking the key disconnects the app immediately.

Costs

You pay the model's usage from your prepaid OceanToken balance, at the prices on the Models page. Ask your agent for an estimate first (estimate_cost). Video jobs place a small hold when they start and charge the actual cost when they finish. Failed jobs are not charged.

Privacy and terms

Prompts and files go to the model you choose, through OceanToken, to produce your result. Generated files are stored in your OceanToken storage and shared as time-limited links. Besides the MCP server at https://mcp.oceantoken.ai/mcp, the setup skill calls OceanToken's own API at https://api.oceantoken.ai with your OCEANTOKEN_API_KEY, only to check the key and to configure your tools to use it. No other service receives data from this plugin. See the privacy policy and terms.

Support

Open an issue, or email [email protected].

Source: README.md at commit 7a58f26

Tools

0
This server asks you to sign in, so SourceWeft could not read its tools without an account and has not indexed them. What a connected account sees can depend on that account and its permissions.

Version history

2
  1. v1.0.1LatestOct 20, 2026
  2. v1.0.0Sep 30, 2026

Score

85% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 daysRequired25 pts
  • Auth declaredRequired10 pts
  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts