v0.5.0

Getting Started

Seren MCP exposes the current tool surface for Seren projects, branches, databases, publishers, payments, managed agents, and seren-cloud operations over the Model Context Protocol (MCP).

POST /mcp — Send JSON-RPC messages GET /mcp — Establish SSE stream (with session) DELETE /mcp — Close session
122 tools in this build
16 documentation sections
Hosted OAuth + local API key workflows

The inventory below is generated directly from the current Rust tool registrations at build time, so the page tracks the live server surface instead of a hand-maintained list.

Server Setup Hosted connection, local CLI startup, auth, and environment variables. Repository Source for the MCP server, CLI, OpenAPI specs, and supporting docs.

For operator workflows, start with get_cloud_overview, list_cloud_agents, list_pending_cloud_approvals, and the managed-agent tools before drilling into individual runs.

Available Tools

Projects

Tool Description
list_projects List all Seren projects accessible to the authenticated user
describe_project Get detailed information about a specific project
create_project Create a new Seren project
delete_project Delete a Seren project
update_project Update a project's settings including name, security options, and compute defaults

Branches

Tool Description
create_branch Create a new branch in a project
delete_branch Delete a branch
list_branches List branches for a project
describe_branch Describe a branch
rename_branch Rename a branch
set_default_branch Set a branch as the default branch for the project
reset_branch Reset a branch to its parent's latest state (destroys all data on the branch)
set_branch_expiration Set or remove branch expiration date

Databases

Tool Description
list_databases List all databases in a branch
list_all_databases List all databases across all projects. Returns a flat list with project and branch names for easy identification.
create_database Create a new database in a branch
get_database_tables List tables in a database schema
describe_table_schema Get schema information for a table
get_database Get details about a specific database
delete_database Delete a database from a branch

SQL

Tool Description
get_connection_string Get connection string for a branch
run_sql Execute a query against a database (SQL for SQL publishers)
run_sql_transaction Execute multiple SQL statements in a single transaction
explain_sql_statement Explain a SQL statement (FORMAT JSON)

Roles

Tool Description
list_roles List all roles in a branch
create_role Create a new database role on a branch
delete_role Delete a database role from a branch
reset_role_password Reset a database role's password, generating a new secure password
reveal_role_password Reveal the current password for a database role

Endpoints

Tool Description
list_endpoints List all endpoints for a branch
create_endpoint Create a new endpoint for a branch
delete_endpoint Delete an endpoint
start_endpoint Start a suspended endpoint
suspend_endpoint Suspend an endpoint
restart_endpoint Restart an endpoint (rolling restart via Kubernetes)
update_endpoint Update an endpoint's settings including autoscaling and suspend timeout
get_endpoint_status Get the current status of an endpoint (running, suspended, etc.)

Organizations & Access

Tool Description
list_organizations List organizations accessible to the authenticated user
list_api_keys List all API keys for an organization
create_api_key Create a new API key for an organization
revoke_api_key Revoke an API key
list_org_oauth_providers List OAuth providers configured for an organization. These are BYOC (Bring Your Own Credentials) OAuth configurations that can be linked to publishers.
get_org_oauth_provider Get details about a specific OAuth provider configuration for an organization.
create_org_oauth_provider Create a new OAuth provider configuration for an organization. This enables BYOC (Bring Your Own Credentials) authentication for publishers.
update_org_oauth_provider Update an OAuth provider configuration for an organization.
delete_org_oauth_provider Delete an OAuth provider configuration from an organization. Warning: This will break any publishers using this OAuth provider.

Agent Store & Publishers

Tool Description
list_agent_publishers List active publishers in the agent store (compact by default). Set verbose=true for full publisher objects. For task-specific recommendations, use suggest_for_task instead.
get_agent_publisher Get details about a specific publisher including pricing info by slug
suggest_for_task Get publisher and agent recommendations for a task. Call this BEFORE using WebSearch/WebFetch to check if a Seren publisher can do the task better. Examples: 'scrape website' returns Firecrawl, 'research topic' returns Perplexity, 'AI search' returns Perplexity. Note: Agent templates are coming soon.
estimate_query_cost Estimate the cost of a publisher query payload without executing it
call_publisher Call a publisher to execute queries, API requests, or MCP operations. This unified tool automatically routes based on parameters: - DATABASE publishers: provide `query` (SQL) and optionally `database` - API publishers: provide `method`, `path`, `headers`, `body` - MCP publishers: provide `tool` + `tool_args` OR `resource_uri` Examples: - Database: call_publisher(publisher: "my-db", query: "SELECT * FROM users") - API: call_publisher(publisher: "firecrawl", method: "POST", path: "/scrape", body: {url: "..."}) - MCP tool: call_publisher(publisher: "my-mcp", tool: "search", tool_args: {query: "..."}) - MCP resource: call_publisher(publisher: "my-mcp", resource_uri: "file:///data.json")
create_publisher Create a new publisher in the agent store. Publishers provide databases or APIs that AI agents can query with micropayments. Requires API key authentication (organization-level).
update_publisher Update an existing publisher's details. Supports both API key and OAuth authentication. Only fields provided will be updated.
update_publisher_pricing Update a publisher's pricing configuration. This updates prices for API calls, database queries, and other billable operations. Requires API key authentication (organization-level). Use get_agent_publisher first to see current pricing.
upload_publisher_logo Upload a logo image for a publisher. Accepts base64-encoded PNG, JPEG, WebP, or SVG images. Maximum size 100KB, automatically resized to 200x200 if larger. Supports both API key and OAuth authentication.
list_agent_templates List available agent templates in the catalog. Templates are executable code (Python, TypeScript, JavaScript) that can be invoked via micropayments.
get_agent_template Get details about a specific agent template by slug
invoke_agent_template Invoke an agent template with input data. Uses SerenBucks when authenticated; otherwise pays via x402 using your configured local wallet. Templates execute code in a sandboxed environment and return structured output.
run_agent_cloud Run an agent task via the unified publisher proxy. Sends a message to a remote A2A agent publisher and returns 202 with the created task. Use get_agent_task to poll for completion.
list_agent_tasks List agent tasks for an organization. Returns tasks ordered by creation time (newest first) with pagination support.
get_agent_task Get details of a specific agent task including status, output, cost breakdown, and A2A protocol metadata.
cancel_agent_task Cancel a running agent task. Releases any billing reservation and stops execution. Only works on tasks that are not yet in a terminal state (completed, failed, or already cancelled).

Payments & Wallets

Tool Description
get_prepaid_balance Get your SerenBucks balance. SerenBucks are credits used to pay for API calls and database queries.
get_wallet_status Get your wallet balance and SerenBucks funding breakdown.
create_prepaid_deposit Deposit SerenBucks with a credit card via Stripe. Returns a checkout URL to complete payment. After payment, SerenBucks are automatically added to your balance.
get_transaction_history Get transaction history for your wallet (deposits, charges, refunds)

Local Wallet & x402

Tool Description
get_local_wallet_address Get the local wallet address. Only available when running seren-mcp locally with WALLET_PRIVATE_KEY environment variable set. Returns the EVM wallet address derived from the private key.
has_local_wallet Check if a local wallet is configured. Returns true if WALLET_PRIVATE_KEY is set and a valid wallet is loaded.
get_onchain_wallet_status Get the configured local wallet's on-chain status on Base, including address, ETH balance, and USDC balance.
get_x402_deposit_requirements Get x402 on-chain deposit requirements for depositing USDC to a publisher. Returns EIP-712 typed data that an agent with an Ethereum wallet must sign to complete the deposit. This is for agents with real crypto wallets; for fiat deposits use create_prepaid_deposit instead.
get_supported Get supported payment protocols and configuration. Returns x402 protocol details including supported payment kinds, networks, and facilitator information for payment protocol discovery.

MCP Publishers

Tool Description
list_mcp_tools List tools available on an MCP publisher. MCP publishers expose tools, resources, and prompts that can be invoked. Use this to discover what capabilities an MCP publisher provides.
list_mcp_resources List resources available on an MCP publisher. MCP publishers can expose resources (like files, data sources) that can be read. Use this to discover what resources an MCP publisher provides.

Managed Agents

Tool Description
get_seren_agent_deployment Get the resolved managed deployment detail for a seren-agent deployment, including the saved prompt, template, resolved tool presets, allowed publisher operations, remote A2A delegation allowlist, runtime overrides, visible config, and secret key names.
list_seren_agent_deployment_revisions List immutable revision snapshots for a managed seren-agent deployment so callers can inspect version history before rollback.
preview_seren_agent_deployment_update Preview an existing managed seren-agent deployment update before applying it. This returns the current resolved managed spec, the proposed resolved spec, and the changed fields so callers can inspect the diff before mutation, including eval-gate changes.
preview_seren_agent_deployment_rollback Preview rolling a managed seren-agent deployment back to a prior revision. This returns the target revision metadata plus the resolved diff against the current deployment.
update_seren_agent_deployment Update an existing managed seren-agent deployment. Use prompt plus optional agent_style, capabilities, access_mode, and performance_profile for the common path, or send the advanced fields when you need eval gates, remote delegation, fallback models, config, secrets, or timeouts.
rollback_seren_agent_deployment Roll a managed seren-agent deployment back to a prior revision. Use the preview tool first to inspect the resolved diff before applying the rollback.
deploy_seren_agent Deploy a managed prompt-based agent through the first-class seren-agent publisher. The simplest inputs are name, mode, prompt, plus optional agent_style, capabilities, access_mode, and performance_profile. Advanced fields still exist for eval gates, remote delegation, fallback models, config, secrets, and timeouts. This path is AWS-first and hides raw runtime internals.

Cloud Environments

Tool Description
list_cloud_environments List reusable cloud deployment environments in the current organization.
get_cloud_environment Get details for a reusable cloud deployment environment.
create_cloud_environment Create a reusable cloud deployment environment.
update_cloud_environment Update a reusable cloud deployment environment.
delete_cloud_environment Delete a reusable cloud deployment environment.

Cloud Deployments

Tool Description
deploy_cloud_agent Deploy a code bundle to the seren-cloud publisher for managed hosting. Supports script and LLM orchestration, always_on (persistent), cron (scheduled), and job modes. Leave compute_backend/runtime_kind unset, or set them to auto, for AWS-first bundle-based routing. Set compute_backend explicitly to force cloudflare_worker or daytona. Auto-routing inspects the uploaded scripts bundle itself: Python/JS/TS entrypoints, shell scripts, Linux binaries, standalone .wasm modules, and Worker JS+.wasm artifacts are all detected from files rather than SKILL.md prose.
list_cloud_agents List cloud agent deployments in the current organization.
get_cloud_overview Get an organization-wide seren-cloud overview with deployment counts, recent runs, and pending approvals.
cloud_agent_status Get status and details of a cloud agent deployment.
start_cloud_agent Start a stopped always-on cloud agent.
stop_cloud_agent Stop a running always-on cloud agent.
run_cloud_agent Trigger a run for a cloud agent. For always_on deployments this proxies the request to the running service; for cron/ephemeral deployments this enqueues a one-shot execution.
cloud_agent_logs Get logs from a running cloud agent's pods.
destroy_cloud_agent Destroy a cloud agent deployment and clean up all K8s resources. This is irreversible.
update_cloud_agent_config Update config, secrets, alert_policy, network_policy, and/or the deployment eval gate for a cloud agent without redeploying code. Provide JSON objects for config/secrets/alert_policy/network_policy, eval_gate_set_id plus eval_gate_max_age_seconds, or the clear_* flags.

Cloud Runs & Approvals

Tool Description
list_cloud_agent_runs List run history for a cloud agent deployment. Supports filters: status, compute_backend, source, has_artifacts, started_after, started_before, and q.
get_cloud_agent_run Get details of a specific run event for a cloud agent deployment, including output, structured events, trace_context, output provenance, and eval_capture trajectory summaries.
list_cloud_agent_run_artifacts List artifacts emitted by a specific run event for a cloud agent deployment.
cancel_cloud_agent_run Cancel a queued/running run event for a cloud agent deployment.
list_all_cloud_runs List all runs across all cloud agent deployments in the organization. Supports filters: status, compute_backend, source, has_artifacts, started_after, started_before, and q.
list_pending_cloud_approvals List all seren-cloud runs currently awaiting approval across the organization.
list_deployment_pending_cloud_approvals List seren-cloud runs currently awaiting approval for a specific deployment.
get_cloud_run_pending_approvals Get the current pending approvals for a seren-cloud run by run ID.
approve_cloud_run_pending_approvals Approve all current pending approvals for a seren-cloud run by run ID and resume it.
reject_cloud_run_pending_approvals Reject all current pending approvals for a seren-cloud run by run ID and resume it.
get_deployment_cloud_run_pending_approvals Get the current pending approvals for a seren-cloud run within a deployment.
get_cloud_run_by_id Get details of a run event by run ID (global path, no deployment ID required), including trace_context, output provenance, and eval_capture trajectory summaries.
compare_cloud_runs Compare eval_capture summaries and replay artifacts for two cloud runs by run ID. Useful for regression checks after rerunning a deployment or managed agent.
list_cloud_run_artifacts List artifacts emitted by a run event using the global run path (no deployment ID required).
cancel_cloud_run_by_id Cancel a queued/running run event by run ID (global path, no deployment ID required).

Cloud Evals

Tool Description
list_cloud_eval_sets List durable eval sets for seren-cloud runs. Optionally scope the list to a single deployment.
create_cloud_eval_set Create a durable eval set for seren-cloud runs. Use deployment_id to scope the set to a specific deployment, criteria to define pass/fail thresholds, metadata for labels or ownership data, and schedule_cron plus schedule_timezone to run the eval set automatically.
get_cloud_eval_set Get a single eval set by ID.
update_cloud_eval_set Replace an existing eval set. Use clear_deployment to remove deployment scoping, or clear_schedule to disable scheduled execution.
list_cloud_eval_cases List eval cases within an eval set.
get_cloud_eval_case Get a single eval case within an eval set.
promote_cloud_run_to_eval_case Promote a terminal seren-cloud run into a durable eval case inside an eval set. The run must already have replay and eval capture data.
create_cloud_eval_run Execute an eval set against a deployment and create durable per-case results.
list_cloud_eval_runs List eval runs within an eval set.
get_cloud_eval_run Get a single eval run by eval set ID and eval run ID.
list_cloud_eval_run_results List per-case results for an eval run.
get_cloud_eval_case_result Get a single per-case result from an eval run.