Create Agent

curl -X POST 'https://app.memorylake.ai/openapi/memorylake/api/v3/agents' \
  -H 'Authorization: Bearer sk_xxxxxx' \
  -H 'Content-Type: application/json' \
  -d '{
    "name": "Support Assistant",
    "custom_id": "support-assistant-v1",
    "description": "Handles customer support inquiries using knowledge base",
    "metadata": {
      "team": "customer-success",
      "tier": "production"
    },
    "model": "gpt-4o",
    "capabilities": {"web_search": {"open_datasets": ["research/academic"]}},
    "policies": {
      "max_turns": 20,
      "max_tool_calls": 10,
      "allow_tools": ["search_kb", "create_ticket"],
      "subagent_timeout": 300,
      "subagent_max_concurrency": 3
    },
    "output": {
      "mode": "text"
    },
    "subagents": [
      {
        "name": "ticket-creator",
        "mode": "fork",
        "context": "Create support tickets from conversation context",
        "inherit_tools": false
      }
    ],
    "system_prompt": "You are a helpful support assistant. Use the knowledge base to answer questions accurately.",
    "model_settings": {
      "temperature": 0.7,
      "max_tokens": 4096
    },
    "runtime_bindings": {
      "knowledge_base": "kb_prod_docs"
    }
  }'

{
  "success": true,
  "data": {
    "id": "agent-f8bf71ab17ed8eccad8c0e8ae708273f",
    "name": "Support Assistant",
    "description": "Handles customer support inquiries using knowledge base",
    "metadata": {
      "team": "customer-success",
      "tier": "production"
    },
    "version": 1,
    "model": "gpt-4o",
    "capabilities": {"web_search": {"max_results": 5, "search_depth": "basic", "chunks_per_source": 2, "open_datasets": ["research/academic"]}},
    "policies": {
      "max_turns": 20,
      "max_tool_calls": 10,
      "allow_tools": ["search_kb", "create_ticket"],
      "deny_tools": [],
      "subagent_timeout": 300,
      "subagent_max_concurrency": 3
    },
    "output": {
      "mode": "text",
      "json_schema": null
    },
    "subagents": [
      {
        "name": "ticket-creator",
        "mode": "fork",
        "context": "Create support tickets from conversation context",
        "inherit_tools": false
      }
    ],
    "skills": [],
    "custom_id": "support-assistant-v1",
    "actor_id": "actor-9ad7a492416c6a8197f42a64b0a4a2d2",
    "latest_version": 1,
    "system_prompt": "You are a helpful support assistant. Use the knowledge base to answer questions accurately.",
    "model_settings": {
      "temperature": 0.7,
      "max_tokens": 4096
    },
    "runtime_bindings": {
      "knowledge_base": "kb_prod_docs"
    },
    "created_at": "2025-03-10T09:00:00Z",
    "updated_at": "2025-03-10T09:00:00Z"
  }
}

curl -X POST 'https://app.memorylake.ai/openapi/memorylake/api/v3/agents' \
  -H 'Authorization: Bearer sk_xxxxxx' \
  -H 'Content-Type: application/json' \
  -d '{
    "name": "Support Assistant",
    "custom_id": "support-assistant-v1",
    "description": "Handles customer support inquiries using knowledge base",
    "metadata": {
      "team": "customer-success",
      "tier": "production"
    },
    "model": "gpt-4o",
    "capabilities": {"web_search": {"open_datasets": ["research/academic"]}},
    "policies": {
      "max_turns": 20,
      "max_tool_calls": 10,
      "allow_tools": ["search_kb", "create_ticket"],
      "subagent_timeout": 300,
      "subagent_max_concurrency": 3
    },
    "output": {
      "mode": "text"
    },
    "subagents": [
      {
        "name": "ticket-creator",
        "mode": "fork",
        "context": "Create support tickets from conversation context",
        "inherit_tools": false
      }
    ],
    "system_prompt": "You are a helpful support assistant. Use the knowledge base to answer questions accurately.",
    "model_settings": {
      "temperature": 0.7,
      "max_tokens": 4096
    },
    "runtime_bindings": {
      "knowledge_base": "kb_prod_docs"
    }
  }'

{
  "success": true,
  "data": {
    "id": "agent-f8bf71ab17ed8eccad8c0e8ae708273f",
    "name": "Support Assistant",
    "description": "Handles customer support inquiries using knowledge base",
    "metadata": {
      "team": "customer-success",
      "tier": "production"
    },
    "version": 1,
    "model": "gpt-4o",
    "capabilities": {"web_search": {"max_results": 5, "search_depth": "basic", "chunks_per_source": 2, "open_datasets": ["research/academic"]}},
    "policies": {
      "max_turns": 20,
      "max_tool_calls": 10,
      "allow_tools": ["search_kb", "create_ticket"],
      "deny_tools": [],
      "subagent_timeout": 300,
      "subagent_max_concurrency": 3
    },
    "output": {
      "mode": "text",
      "json_schema": null
    },
    "subagents": [
      {
        "name": "ticket-creator",
        "mode": "fork",
        "context": "Create support tickets from conversation context",
        "inherit_tools": false
      }
    ],
    "skills": [],
    "custom_id": "support-assistant-v1",
    "actor_id": "actor-9ad7a492416c6a8197f42a64b0a4a2d2",
    "latest_version": 1,
    "system_prompt": "You are a helpful support assistant. Use the knowledge base to answer questions accurately.",
    "model_settings": {
      "temperature": 0.7,
      "max_tokens": 4096
    },
    "runtime_bindings": {
      "knowledge_base": "kb_prod_docs"
    },
    "created_at": "2025-03-10T09:00:00Z",
    "updated_at": "2025-03-10T09:00:00Z"
  }
}

POST /openapi/memorylake/api/v3/agents

Creates a new agent with the specified configuration. You provide a name and custom identifier, and optionally attach a description, model, capabilities, policies, subagents, and other settings. The agent is created at version 1.

Permission required: agent:create

Request Body

string

required

Agent name

string

required

Your own identifier for this agent — the one it already has in your system. Must be unique within your tenant. Pass by_custom_id=true on later requests to address the agent by it instead of by id. See Your own identifiers.

string

Agent description

string

Whether the platform hosts the agent’s runtime. INTERNAL (the default) runs the agent here and supports versions, forking and A2A. EXTERNAL registers memory only — the agent runs on your side, so every runtime field below (system_prompt, model, model_settings, capabilities, policies, output, runtime_bindings, subagents, skills) is ignored.

object

Arbitrary key-value pairs for storing additional agent information

string

Model to use for the agent (e.g. gpt-4o, claude-sonnet-4-20250514)

object

Capabilities the agent has in every conversation, keyed by capability name with each capability’s configuration as the value. You can set web_search and web_fetch here, in the same form as capabilitiesOverride in an A2A message; pii and vega can only be turned on per message. A name this field doesn’t recognize is dropped without an error. Omit the field and the agent has neither.

object

Execution policies that constrain agent behavior

Show Policies object

integer

Maximum number of conversation turns per run

integer

Maximum number of tool calls per turn

array

Allowlist of tool names the agent may invoke

array

Denylist of tool names the agent must not invoke

integer

Maximum seconds a subagent may run before being cancelled

integer

Maximum number of subagents that may run concurrently

object

Output configuration for the agent

Show Output object

string

Output mode (text or json)

object

JSON Schema that the agent’s output must conform to when mode is json

array

Subagent definitions to attach to this agent

Show Subagent object

string

Subagent name

string

Execution mode: fork or swarm

string

Context instructions passed to the subagent

boolean

Whether the subagent inherits the parent agent’s tools

array

Skills to attach to this agent. Every skill must be usable: if any has not passed security review, does not exist, or has no package, the request fails with 400 and SKILL_NOT_USABLE, naming each skill and the reason.

Show Skill reference

string

Skill ID

integer

Pin the reference to this version. Omit to follow the latest version. A pinned reference is checked against that version’s own review result.

string

System prompt used to instruct the agent

object

Model-specific settings such as temperature and max tokens

object

Runtime binding configuration for tools and data sources

curl -X POST 'https://app.memorylake.ai/openapi/memorylake/api/v3/agents' \
  -H 'Authorization: Bearer sk_xxxxxx' \
  -H 'Content-Type: application/json' \
  -d '{
    "name": "Support Assistant",
    "custom_id": "support-assistant-v1",
    "description": "Handles customer support inquiries using knowledge base",
    "metadata": {
      "team": "customer-success",
      "tier": "production"
    },
    "model": "gpt-4o",
    "capabilities": {"web_search": {"open_datasets": ["research/academic"]}},
    "policies": {
      "max_turns": 20,
      "max_tool_calls": 10,
      "allow_tools": ["search_kb", "create_ticket"],
      "subagent_timeout": 300,
      "subagent_max_concurrency": 3
    },
    "output": {
      "mode": "text"
    },
    "subagents": [
      {
        "name": "ticket-creator",
        "mode": "fork",
        "context": "Create support tickets from conversation context",
        "inherit_tools": false
      }
    ],
    "system_prompt": "You are a helpful support assistant. Use the knowledge base to answer questions accurately.",
    "model_settings": {
      "temperature": 0.7,
      "max_tokens": 4096
    },
    "runtime_bindings": {
      "knowledge_base": "kb_prod_docs"
    }
  }'

Response

object

Show Agent object

string

Unique agent identifier

string

Agent name

string

Agent description

object

Arbitrary key-value pairs for storing additional agent information

integer

Current active version number

string

Model used by the agent

object

Capabilities the agent has, keyed by capability name. The platform fills in default values for fields you didn’t set, so a configuration can include fields these docs don’t describe.

object

Execution policies that constrain agent behavior

Show Policies object

integer

Maximum number of conversation turns per run

integer

Maximum number of tool calls per turn

array

Allowlist of tool names the agent may invoke

array

Denylist of tool names the agent must not invoke

integer

Maximum seconds a subagent may run before being cancelled

integer

Maximum number of subagents that may run concurrently

object

Output configuration for the agent

Show Output object

string

Output mode

object

JSON Schema for structured output

array

Subagent definitions attached to this agent

Show Subagent object

string

Subagent name

string

Execution mode: fork or swarm

string

Context instructions passed to the subagent

boolean

Whether the subagent inherits the parent agent’s tools

array

Skills attached to this agent

Show Skill reference

string

Skill identifier

integer

Skill version number

string

Custom identifier for external system integration

string

Actor identifier associated with this agent

integer

Latest version number available for this agent

string

System prompt used to instruct the agent

object

Model-specific settings such as temperature and max tokens

object

Runtime binding configuration for tools and data sources

string

Creation timestamp

string

Last update timestamp

{
  "success": true,
  "data": {
    "id": "agent-f8bf71ab17ed8eccad8c0e8ae708273f",
    "name": "Support Assistant",
    "description": "Handles customer support inquiries using knowledge base",
    "metadata": {
      "team": "customer-success",
      "tier": "production"
    },
    "version": 1,
    "model": "gpt-4o",
    "capabilities": {"web_search": {"max_results": 5, "search_depth": "basic", "chunks_per_source": 2, "open_datasets": ["research/academic"]}},
    "policies": {
      "max_turns": 20,
      "max_tool_calls": 10,
      "allow_tools": ["search_kb", "create_ticket"],
      "deny_tools": [],
      "subagent_timeout": 300,
      "subagent_max_concurrency": 3
    },
    "output": {
      "mode": "text",
      "json_schema": null
    },
    "subagents": [
      {
        "name": "ticket-creator",
        "mode": "fork",
        "context": "Create support tickets from conversation context",
        "inherit_tools": false
      }
    ],
    "skills": [],
    "custom_id": "support-assistant-v1",
    "actor_id": "actor-9ad7a492416c6a8197f42a64b0a4a2d2",
    "latest_version": 1,
    "system_prompt": "You are a helpful support assistant. Use the knowledge base to answer questions accurately.",
    "model_settings": {
      "temperature": 0.7,
      "max_tokens": 4096
    },
    "runtime_bindings": {
      "knowledge_base": "kb_prod_docs"
    },
    "created_at": "2025-03-10T09:00:00Z",
    "updated_at": "2025-03-10T09:00:00Z"
  }
}

MemoryLake API

Memory Router API

Request Body

Response

​Request Body

​Response

Request Body

Response