Can AI agents automate Toggl 2.0?

Toggl 2.0 is automatable today through the official CLI. This record is Toggl 2.0 workspaces (tasks, projects, time entries, time blocks, reports, and organization members). It is not Toggl Track. The CLI package is @togglhq/cli. `toggl auth` opens a browser for OAuth and stores the session locally. Commands cover projects, tasks, time entries, time blocks, and users, and `--json` prints machine-readable output. The same surface is available to MCP clients through the local package @togglhq/mcp, which also signs in through the browser on first use. There is no separate MCP or CLI request cap; access follows the Toggl 2.0 plan, including a 30 report-data requests per month limit on Free. The public HTTPS API is at https://focus.toggl.com/api and accepts JSON. The authentication guide uses Authorization: Bearer with a toggl_sk_ API key, one active key per user. The tasks reference samples instead show HTTP basic auth with email and password. Hourly quotas are 30 requests on Free, 240 on Starter, and 600 on Premium, per user per organization, and an exhausted quota returns HTTP 402. The Free plan is $0. Starter is $9 per license per month and Premium is $16. Pricing lists web, Windows, macOS, Android, and iOS clients. Connector catalogs, webhooks, and an llms.txt index were not established. Computer-use viability is unassessed. The CLI and MCP both score 7. The CLI is the best path because `--json` commands do not require an MCP client after `toggl auth`. The API scores 6 because of the hourly plan quota and the split authentication examples.

Best path today: cli · Overall automatability: 7/10