REST API Documentation

List incidents with pagination and optional Unix timestamp filters (page, per_page, from, to)

Get a specific incident by UUID, including execution state and logs

Create and start a new API-driven incident. Request body requires task and may include optional context.

Submit thumbs-up / thumbs-down style execution feedback for a completed incident so Akmatori can store operator signal alongside the run.

List all available skills

Create a new skill

Get skill details by name

Update skill configuration

Delete a skill (user-created only)

Get skill prompt content

Update skill prompt

List tools assigned to a skill

Assign tool instances to a skill

List skill scripts

Clear all scripts for a skill

Get a specific script by filename

Create or update a skill script

Delete a specific script

Sync skills from filesystem to database

List available tool types

List configured tool instances

Create a new tool instance

Get tool instance details

Update tool configuration

Delete a tool instance

List SSH keys for an SSH tool

Add an SSH key to a tool

Update an SSH key

Delete an SSH key

List supported alert source types

List configured alert sources

Create a new webhook alert source. Deprecated source types such as slack_channel are rejected; configure Slack listeners through Channels instead.

Get alert source details by UUID

Update alert source configuration

Delete an alert source

List messaging provider integrations such as Slack, with credentials masked in responses.

Create a provider integration with provider, name, optional metadata, and provider-specific credentials.

Get integration details by UUID, including masked credential metadata.

Update integration name, credentials, metadata, or enabled state.

Delete a messaging integration and its provider configuration.

List provider channels with optional filters for integration, posting, listening, default post target, and enabled state.

Create a channel for a messaging integration using integration_uuid, provider external_id, posting/listening flags, and optional extraction prompt.

Get channel details by UUID, including its parent integration when loaded.

Update channel routing, display name, listen/post flags, extraction prompt, human-message processing, or enabled state.

Delete a channel mapping and reload listener routing for supported providers.

List agent-only scheduled jobs, including channel target, next run time, system flag, last-run status, and each job's secret-free tool allowlist.

Create an agent cron job with name, five-field cron schedule, prompt, optional channel_uuid, enabled, and tool_instance_ids.

Get a scheduled job, including last-run status, next run time, delivery target, system flag, and assigned tool summaries.

Update schedule, prompt, channel target, enabled state, or replace the per-cron tool_instance_ids allowlist. Omit tool_instance_ids to keep existing tools; pass [] to clear them.

Start a job immediately in the background and return 202 Accepted.

Delete an operator-created scheduled job. Seeded system jobs can be disabled but return 409 Conflict on delete.

List all context files

Upload a new context file via multipart form with file and filename, plus optional description.

Get context file metadata

Download context file content

Delete a context file

Validate context file references in text

List all runbooks for AI-guided incident investigation

Create a new runbook (title, content). Akmatori syncs the markdown file automatically and, when QMD is configured, triggers background re-indexing.

Get runbook details by ID

Update runbook title and content, then resync exported markdown and optional QMD indexing metadata

Delete a runbook and clean up its synced markdown export

List stored cross-incident memories, with optional filtering by scope, source, or query terms.

Create a new memory record that can be reused across future incidents.

List available memory scopes so operators can understand where shared context will be applied.

Get a single memory entry by ID, including its scope and extracted content.

Update an existing memory entry when shared knowledge needs correction or enrichment.

Delete a memory entry that should no longer influence incident analysis.

List all HTTP connectors for external API integration

Create a new HTTP connector with tool_type_name, base_url_field, optional auth_config, and generated tools definitions.

Get HTTP connector details by ID

Update HTTP connector configuration

Delete an HTTP connector

List all MCP server configurations for external tool proxying

Register a new MCP server with name, transport, and namespace_prefix, using either url or command/args/env_vars.

Get MCP server configuration by ID

Update MCP server configuration

Delete an MCP server registration

Get general settings (base URL for links in notifications)

Update general settings

Legacy Slack settings endpoint. Returns 410 Gone; use /integrations and /channels instead.

Legacy Slack settings endpoint. Returns 410 Gone; configure Slack through messaging integrations and channels.

List all saved LLM configurations and return the currently active config ID

Create a new LLM configuration for OpenAI, Anthropic, Google, OpenRouter, or a custom provider, including optional thinking_level (off, minimal, low, medium, high, xhigh).

Get a single LLM configuration by numeric ID

Update an existing LLM configuration, including name, model, thinking level, base URL, or API key

Delete an LLM configuration that is no longer needed

Mark a specific LLM configuration as the active one used for incident execution

Get outbound proxy configuration, including the masked proxy URL, no_proxy, and per-service toggles for llm, slack, zabbix, victoria_metrics, catchpoint, grafana, pagerduty, netbox, kubernetes, and jira. SSH is currently reported as unsupported.

Update proxy configuration, including the shared proxy URL, no_proxy, and service-specific enablement flags used by the agent worker and supported integrations.

Get incident data retention policy settings, including enablement, retention window, and cleanup interval

Update retention policy settings (enabled, retention_days, cleanup_interval_hours)

Get global response-formatting settings used to rewrite final agent answers before persistence, including the editable output_schema_example used to infer the final JSON shape.

Update formatting settings (enabled, system_prompt, max_tokens, temperature, output_schema_example). Empty prompt or schema values fall back to built-in defaults; custom schema examples must be valid JSON objects and stay within the 8 KB limit.

Exchange credentials for a JWT token

Verify the current JWT token and return the authenticated user context

Check whether the instance still requires initial admin setup

Create the initial admin account during first-run bootstrap

Liveness and readiness check for Akmatori services

Receive alerts from monitoring systems (no auth required)

Interactive Swagger UI for exploring and testing API endpoints

OpenAPI 3.1 specification in YAML format