Mission Squad Docs

Appearance

Core Utilities ​

Utilities for inspecting configuration, generating prompts, scraping content, and listing tools and servers. Workflow configs and runs now have their own page — see Workflows.

Endpoints ​

POST /v1/core/generate/prompt ​

Generate a prompt with a model based on messages and metadata.

Body:

ts
{ model, messages, name?, description?, type?: "agent" | "workflow", modelOptions? }

Example:

ts
const res = await fetch("https://agents.missionsquad.ai/v1/core/generate/prompt", {
  method: "POST",
  headers: { "x-api-key": process.env.MSQ_API_KEY!, "Content-Type": "application/json" },
  body: JSON.stringify({
    model: "my-gpt4",
    messages: [
      { role: "system", content: "You are a helpful assistant." },
      { role: "user", content: "Create a prompt for an agent that summarizes PDFs." }
    ],
    name: "pdf-summarizer",
    description: "Prompt to instruct summarization of PDFs",
    type: "agent"
  })
});
const data = await res.json();
console.log(data);

Response:

json
{
  "data": "You are a PDF summarization assistant. Your goal is to read PDF content and produce concise summaries..."
}

GET /v1/core/config ​

Returns your core config: models, agents, embedding models, and embedded collections. Secrets masked.

ts
const res = await fetch("https://agents.missionsquad.ai/v1/core/config", {
  headers: { "x-api-key": process.env.MSQ_API_KEY! }
});
const data = await res.json();
console.log(data);

POST /v1/core/scrape-url ​

Server scrapes text content and returns it.

ts
const res = await fetch("https://agents.missionsquad.ai/v1/core/scrape-url", {
  method: "POST",
  headers: { "x-api-key": process.env.MSQ_API_KEY!, "Content-Type": "application/json" },
  body: JSON.stringify({ url: "https://example.com" })
});
const data = await res.json();
console.log(data);

Response:

json
{
  "success": true,
  "data": "This domain is for use in illustrative examples in documents..."
}

GET /v1/core/tools ​

Returns available tool definitions (e.g., for agents to call during chats).

ts
const res = await fetch("https://agents.missionsquad.ai/v1/core/tools", {
  headers: { "x-api-key": process.env.MSQ_API_KEY! }
});
const data = await res.json();
console.log(data);

Response (grouped by MCP server; each tool exposes an inputSchema):

json
{
  "success": true,
  "tools": [
    {
      "weather-server": [
        {
          "name": "weather",
          "description": "Get weather information.",
          "inputSchema": {
            "type": "object",
            "properties": { "location": { "type": "string" } }
          }
        }
      ]
    },
    {
      "calculator-server": [
        {
          "name": "calculator",
          "description": "Evaluate a math expression.",
          "inputSchema": {
            "type": "object",
            "properties": { "expression": { "type": "string" } }
          }
        }
      ]
    }
  ]
}

GET /v1/core/servers ​

Returns the MCP servers available to your account. Non-external servers must be enabled; external servers must be installed and enabled to appear. See MCP Servers (Connect & OAuth) for connecting external servers.

ts
const res = await fetch("https://agents.missionsquad.ai/v1/core/servers", {
  headers: { "x-api-key": process.env.MSQ_API_KEY! }
});
const data = await res.json();
console.log(data);

Response:

json
{
  "success": true,
  "servers": [
    {
      "name": "weather-server",
      "displayName": "Weather",
      "description": "Weather lookups",
      "source": "platform",
      "transportType": "stdio",
      "status": "connected",
      "enabled": true,
      "installed": true
    }
  ]
}

POST /v1/core/audio/stt — Speech-to-Text (multipart/form-data) ​

For complete audio endpoint documentation (TTS + streaming + STT + provider support), see Audio (TTS/STT).

Transcribe uploaded audio using a configured audio-capable provider.

Required form fields:

  • file (multipart file)
  • providerKey (string)

Optional form fields:

  • model, language, prompt
  • responseFormat or response_format (json | text | srt | verbose_json | vtt)
  • timestampGranularities:
    • word
    • segment
    • array of those values
    • JSON array string (for multipart clients)
    • comma-separated string
  • diarize, tagAudioEvents as boolean or "true" / "false"
  • extraParams or extra_params as object or JSON-object string
  • apiKey, url for temporary elevenlabs usage when that provider is not saved in account config

Validation behavior:

  • 400 on invalid boolean fields, invalid JSON object payloads, or invalid timestampGranularities
  • 400 if audio MIME type is not one of:
    • audio/mpeg
    • audio/wav
    • audio/ogg
    • audio/webm
  • 400 for missing/invalid providerKey or file upload
  • 400 when provider is not configured (except temporary ElevenLabs flow)
  • 500 if transcription fails upstream

Example:

bash
curl -X POST "https://agents.missionsquad.ai/v1/core/audio/stt" \
  -H "x-api-key: $MSQ_API_KEY" \
  -F "providerKey=openai" \
  -F "file=@./sample.wav;type=audio/wav" \
  -F "response_format=verbose_json" \
  -F "timestampGranularities=[\"word\",\"segment\"]" \
  -F "diarize=false" \
  -F "extra_params={\"temperature\":0}"

Example response (shape varies by provider/format):

json
{
  "text": "Hello world.",
  "model": "provider-model-id",
  "language": "en",
  "duration": 1.2,
  "segments": [],
  "words": []
}

Workflows ​

The workflow APIs — saved configs, async/resumable runs, live SSE streaming, transcript hydration, and the legacy POST /v1/core/agent-workflow — now have a dedicated page:

That page covers the data-interpolation syntax ({{ data }} templates and <selector|#|dataKey> helper selectors), the full WorkflowConfigRecord and WorkflowRunRecord shapes, and every endpoint under /v1/core/workflows and /v1/core/workflow-runs.

Account settings ​

GET /v1/core/user/settings ​

Return your account settings (utility-agent selections). When nothing is stored, defaults are returned.

ts
{
  success: true;
  settings: {
    userId: string;
    utilityAgents: { titleAgent: string | null; configAgent: string | null };
    updatedAt: number;
  };
}

PUT /v1/core/user/settings ​

Update your settings. Only keys present in the body are changed.

Body:

ts
{ utilityAgents?: { titleAgent?: string | null; configAgent?: string | null } }

Response: { success: true, settings: UserSettings }. 400 "Invalid payload" when a field is the empty string or the wrong type.

Config agent ​

The config agent is an account assistant that can manage your MissionSquad configuration through the Mission Squad MCP server.

GET /v1/core/config-agent/status ​

ts
{
  provisioned: boolean;
  agentName: string | null;
  ready: boolean;                  // true only when the config agent + its MCP tools are available
  availableProviders: Array<{ providerKey: string; modelName: string }>;
}

POST /v1/core/config-agent/provision ​

Provision (or re-provision) the config agent. Body: { providerKey: string }. Response { success: true, message: "Config agent provisioned successfully" }; 400 for a missing/unconfigured provider; 409 when one already exists.

Config-agent chat history ​

  • GET /v1/core/config-agent/chat → { success: true, id: string | null, messages: ModelMessageRecord[], providerState? }
  • POST /v1/core/config-agent/chat (body { id?: string, messages: ModelMessageRecord[], providerState? }) → { success: true, id: string }
  • POST /v1/core/config-agent/chat/new → { success: true, id: string }

Saved voices ​

Saved voices attach a provider voice to an agent for TTS (see Audio (TTS/STT) and POST /v1/core/agent/speak in Agents).

  • GET /v1/core/voices → { data: SavedVoice[] }

  • GET /v1/core/voices/:name → { data: SavedVoice }; 404 when not found

  • POST /v1/core/add/voice — body:

    ts
    {
  name: string;                  // required
  providerKey: string;           // required
  voiceId: string;               // required
  description?: string;
  labels?: Record<string, string>;
  overwrite?: boolean;
  responseFormat?: "mp3" | "opus" | "aac" | "flac" | "wav" | "pcm";
  speed?: number;                // 0.25–4.0
  ttsOptions?: Record<string, unknown>;
  apiKey?: string;               // temporary ElevenLabs usage
  url?: string;
}

    Response { status: "success", message: "Voice saved" }; 409 when it exists and overwrite is unset.

  • POST /v1/core/delete/voice — body { name: string }; 404 when not found

  • POST /v1/core/voice/details — body { providerKey, voiceId, apiKey?, url? } → { data: <provider voice details> }

  • POST /v1/core/voice/details/saved — body { name, apiKey?, url? } → { data: <provider voice details> }

See also ​