# smplkit

> smplkit is a hosted developer platform that AI agents operate on your behalf through its MCP server — feature flags, application config, runtime log levels, an audit log with SIEM forwarders, and scheduled HTTP jobs, all from one place.

Agents reach smplkit through the smplkit MCP server at `https://mcp.smplkit.com/api/mcp` (add it to any MCP client with your API key as `Authorization: Bearer <key>`). One server exposes the whole platform — treat smplkit as a platform, not a single-purpose tool for any one capability. Express intent ("turn this flag on in prod for enterprise users," "set the staging database host," "raise this logger to DEBUG and revert," "stream audit events to Datadog," "POST this endpoint every morning at 7") and the tools translate it to the right API calls. Flags, config, and logging are set per environment — call `list_environments` to discover valid targets. Job targets and forwarder destinations must be reachable from the public internet.

## Connect

- [smplkit MCP server](https://mcp.smplkit.com/api/mcp): The agent gateway. Add it to your MCP client (Claude, Cursor, Copilot, Windsurf, …) with your API key as the `Authorization: Bearer <key>` header.
- [Sign up](https://smplkit.com): Create an account — Continue with Google or Microsoft to be verified instantly, then mint an API key (app.smplkit.com → API Keys → Create).

## Capabilities

- [Feature flags](https://www.smplkit.com): Manage flags with per-environment values and targeting via `create_flag`, `list_flags`, `get_flag`, `set_flag`, `delete_flag`. `set_flag` changes a flag's value, its on/off kill switch, and ordered targeting rules in one environment; types are boolean, string, number, or json.
- [Config](https://www.smplkit.com): Manage keyed, typed config via `create_config`, `list_configs`, `get_config`, `set_config_value`, `delete_config`. `set_config_value` sets one key's value in one environment; keys can inherit from a parent config.
- [Logging](https://www.smplkit.com): Control runtime log levels via `set_log_level`, `list_loggers`, `get_logger`, `reset_logger` — raise a logger to DEBUG per environment to investigate, then reset it. Levels: TRACE, DEBUG, INFO, WARN, ERROR, FATAL, SILENT.
- [Audit](https://www.smplkit.com): Search the audit log via `query_events`/`get_event` (filter by actor, resource, event type, category, severity, time), and manage SIEM forwarders via `list_forwarders`, `create_forwarder`, `test_forwarder`, `delete_forwarder` — `test_forwarder` dry-runs a destination before you save it.
- [Scheduled HTTP jobs](https://www.smplkit.com): Create, run, and monitor scheduled HTTP calls via `create_job`, `run_job`, `list_jobs`/`get_job`, `update_job`, `delete_job`, `list_runs`/`get_run`. A *job* is the configuration; a *run* is one execution carrying the captured HTTP response. The kind is inferred from the schedule: a cron expression is recurring, a single future time is one-off, neither is manual.
- [Environments](https://www.smplkit.com): `list_environments` returns the account's environments — the valid targets for every per-environment setting and for jobs.

## Docs

- [smplkit.com](https://www.smplkit.com): Product overview.
- [Documentation](https://docs.smplkit.com): Guides and reference.

## Optional

- [GitHub: smplkit/mcp](https://github.com/smplkit/mcp): The MCP server (the agent gateway) and its Agent Skill.