REST API Reference

• DELETE /mltk/agents/{agent_name} DELETE /mltk/agents/{agent_name} Begin async teardown: calls SCS `DELETE {scs_url_base}/agent/aws_agentcore/{runtime_id}`, sets the agent state to `\"Deleting\"`, and dispatches an `agentstatus` job to monitor cleanup.
• GET /mltk/agents/{agent_name} GET /mltk/agents/{agent_name} Fetch a single agent record.
• GET /mltk/agents GET /mltk/agents List all agents the caller is permitted to read.
• POST /mltk/agents POST /mltk/agents Create a new agent.
• PUT /mltk/agents PUT /mltk/agents Merge updates into an existing agent.
• DELETE /mltk/aicommander/{service}/{connection_name}/{model} DELETE /mltk/aicommander/{service}/{connection_name}/{model} Remove a single model entry.
• GET /mltk/aicommander_metadata GET /mltk/aicommander_metadata Returns provider metadata / supported-provider feature flags.
• GET /mltk/aicommander/{service}/{connection_name}/{model} GET /mltk/aicommander/{service}/{connection_name}/{model} Fetch a single model's settings for editing in the UI.
• GET /mltk/aicommander GET /mltk/aicommander Returns the full set of saved provider configurations as rendered for the AI Commander UI.
• POST /mltk/aicommander_metadata POST /mltk/aicommander_metadata Live-test a provider configuration without saving it.
• POST /mltk/aicommander POST /mltk/aicommander Upsert a provider / connection / model configuration.
• Capabilities Appendix B of the AITK REST API reference: Capabilities.
• KV-store collections Appendix A of the AITK REST API reference: KV-store collections.
• Source code layout Appendix C of the AITK REST API reference: Source code layout.
• DELETE /mltk/container_connection/{container_type}/{connection_name} DELETE /mltk/container_connection/{container_type}/{connection_name} Clears all configuration fields for the stanza (does not remove the stanza itself), effectively disabling it.
• DELETE /mltk/start_container/{model_name} DELETE /mltk/start_container/{model_name} Tear down the container/deployment for `model_name`.
• GET /mltk/container_connection/{container_type}/{connection_name} GET /mltk/container_connection/{container_type}/{connection_name} Fetch one connection.
• GET /mltk/container_connection/{container_type} GET /mltk/container_connection/{container_type} Cluster-wide resource defaults & HPA settings for a container type.
• GET /mltk/container_data GET /mltk/container_data Returns the per-deployment state from `dsdl_container.conf` enriched with a live health probe.
• GET /mltk/container_logs GET /mltk/container_logs Inventory of dev/prod stanzas grouped by mode and active/inactive state.
• GET /mltk/start_container GET /mltk/start_container Summary counts of active / inactive containers and DEV / PROD splits.
• GET /mltk/stop_container GET /mltk/stop_container Returns the catalog of available Docker images from `docker_images.conf`.
• GET /mltk/test_container_connection GET /mltk/test_container_connection Returns the active container connection configuration with secrets decoded.
• POST /mltk/container_connection POST /mltk/container_connection Upsert a container connection stanza.
• POST /mltk/container_logs POST /mltk/container_logs Fetch a container's stdout/stderr logs.
• POST /mltk/start_container POST /mltk/start_container Start a container (Docker) or deployment (Kubernetes).
• POST /mltk/stop_container POST /mltk/stop_container Equivalent to `DELETE /mltk/start_container/{model}` — stops the Docker container or deletes the Kubernetes deployment, clears runtime state.
• POST /mltk/test_container_connection POST /mltk/test_container_connection Same body shape as `POST /mltk/container_connection` plus active connectivity verification.
• DELETE /mltk/experiments/{experiment_id} DELETE /mltk/experiments/{experiment_id} Cascade delete: 1.
• GET /mltk/experiments/{experiment_id}/history GET /mltk/experiments/{experiment_id}/history All history records for the experiment, queried from KV collection `experiment_history` with filter `{\"experimentId\": \"...\"}`.
• GET /mltk/experiments/{experiment_id} GET /mltk/experiments/{experiment_id} Fetch a single experiment.
• GET /mltk/experiments GET /mltk/experiments Splunk-style list response wrapping every experiment stanza in `experiments.conf`.
• POST /mltk/experiments (create) POST /mltk/experiments Create a new experiment.
• POST /mltk/experiments/{experiment_id}/clone_models POST /mltk/experiments/{experiment_id}/clone_models Clone all trained models for the experiment into a different app/name.
• POST /mltk/experiments/{experiment_id}/history POST /mltk/experiments/{experiment_id}/history Append a history record.
• POST /mltk/experiments/{experiment_id} (update) POST /mltk/experiments/{experiment_id} Update an existing experiment.
• GET /mltk/agent_feature_flag GET /mltk/agent_feature_flag Returns the agent-builder feature flag (internally keyed `aitk_agent_builder_feature_enabled`, but the response key is `mltk_hosted_llm` for legacy reasons).
• GET /mltk/ctsm_feature GET /mltk/ctsm_feature Returns the current CTSM opt-out and acknowledge flags.
• GET /mltk/mltk_feature_flags GET /mltk/mltk_feature_flags Returns the hosted-LLM feature flag the UI uses to decide whether Splunk Hosted Models are available.
• POST /mltk/ctsm_acknowledge POST /mltk/ctsm_acknowledge Mark the CTSM notice as acknowledged.
• POST /mltk/ctsm_feature POST /mltk/ctsm_feature Opt out of CTSM.
• DELETE /mltk/mcp_connection/{name} DELETE /mltk/mcp_connection/{name} Delete the connection and its token from password storage.
• GET /mltk/mcp_connection[?name=...] GET /mltk/mcp_connection[?name=...] List all MCP connections, or fetch one by name.
• POST /mltk/mcp_connection/test POST /mltk/mcp_connection/test Test connectivity.
• POST /mltk/mcp_connection/tools POST /mltk/mcp_connection/tools List the tools exposed by a saved or inline MCP connection.
• POST /mltk/mcp_connection POST /mltk/mcp_connection Create a new MCP connection.
• PUT /mltk/mcp_connection PUT /mltk/mcp_connection Update a connection by name.
• GET /mltk/list_connections GET /mltk/list_connections Combined list of LLM, Docker, and Kubernetes connections.
• GET /mltk/list_models GET /mltk/list_models List all MLTK models — both ONNX-uploaded models and SageMaker-registered ones.
• GET /mltk/register_model?model_name=... GET /mltk/register_model?model_name=... Fetch a SageMaker model configuration by name.
• POST /mltk/register_model?test_connection=1 POST /mltk/register_model?test_connection=1 Same body as 7.4 but the handler only probes AWS (`sagemaker:DescribeEndpoint`) — nothing is persisted.
• POST /mltk/register_model POST /mltk/register_model Register a new SageMaker endpoint as an MLTK model.
• POST /mltk/upload_model POST /mltk/upload_model Upload an ONNX model.
• PUT /mltk/register_model PUT /mltk/register_model Update the mutable fields of an existing SageMaker model: `input_feature_map`, `output_prediction_map`, `openapi_spec`, `batch_size`.
• POST /mltk/run_history POST /mltk/run_history Paginate agent run history.
• PUT /mltk/hec_operations PUT /mltk/hec_operations Push one or more AI agent events into the configured Splunk index via `POST /services/receivers/simple` (HEC).
• DELETE /mltk/vector_stores/{connection_name} DELETE /mltk/vector_stores/{connection_name} Remove the configuration and its AWS credentials from password storage.
• GET /mltk/vector_stores GET /mltk/vector_stores List all vector store configurations.
• POST /mltk/vector_stores/test POST /mltk/vector_stores/test Probe an AWS Knowledge Base.
• POST /mltk/vector_stores POST /mltk/vector_stores Create a vector store config.
• PUT /mltk/vector_stores PUT /mltk/vector_stores Update an existing vector store.

No items match the selected filters. Clear filters

/servicesNS/{owner}/Splunk_ML_Toolkit/mltk/<handler>[/<subpath>]

[script:mltk]
match              = /mltk
script             = util/rest_handler.py
scripttype         = persist
requireAuthentication = true
output_modes       = json
passHttpHeaders    = true
passSystemAuth     = true
passPayload        = true
python.version     = python3

[expose:mltk]
pattern = mltk/**
methods = GET, POST, PUT, DELETE