CLI

scheduler0 is a command-line client for the Scheduler0 API, written in Go and distributed as a single binary for macOS, Linux and Windows. Source: github.com/scheduler0/scheduler0-cli.

Installation

Install script (macOS, Linux)

Detects OS and architecture, downloads the latest release, verifies its SHA-256 checksum and installs to /usr/local/bin (using sudo if needed):

curl -fsSL https://raw.githubusercontent.com/scheduler0/scheduler0-cli/main/install.sh | bash

Pin a version or change the install directory:

curl -fsSL https://raw.githubusercontent.com/scheduler0/scheduler0-cli/main/install.sh | VERSION=0.0.4 bash
curl -fsSL https://raw.githubusercontent.com/scheduler0/scheduler0-cli/main/install.sh | INSTALL_DIR=$HOME/.local/bin bash

Release binaries

Archives for every release are on the Releases page, named scheduler0_<version>_<os>_<arch>.tar.gz (.zip on Windows), alongside a checksums.txt with SHA-256 sums.

curl -fsSL -o scheduler0.tar.gz \
  https://github.com/scheduler0/scheduler0-cli/releases/latest/download/scheduler0_0.0.4_darwin_arm64.tar.gz
tar -xzf scheduler0.tar.gz
sudo mv scheduler0 /usr/local/bin/
scheduler0 --version

Replace darwin_arm64 with darwin_amd64, linux_amd64 or linux_arm64 as needed. On Windows, extract scheduler0.exe from the .zip and put it on your PATH.

If macOS Gatekeeper blocks the binary: xattr -d com.apple.quarantine /usr/local/bin/scheduler0.

Docker

Multi-arch images (linux/amd64, linux/arm64) are pushed to Amazon ECR Public on every release, tagged latest and <version>. The image is public.ecr.aws/p2b2u7y5/scheduler0-cli; scheduler0 is the entrypoint, so pass the subcommand as the container arguments.

docker run --rm \
  -v "$HOME/.scheduler0:/root/.scheduler0" \
  public.ecr.aws/p2b2u7y5/scheduler0-cli:latest \
  projects list

Mount ~/.scheduler0 so the container can read the session written by scheduler0 login, or pass SCHEDULER0_* environment variables instead (see CI).

From source

git clone https://github.com/scheduler0/scheduler0-cli.git
cd scheduler0-cli
go build -o scheduler0 .
sudo mv scheduler0 /usr/local/bin/

scheduler0 login

Opens app.scheduler0.com in your browser, where you approve access with your dashboard session. A short-lived credential is written to ~/.scheduler0/config.json and used for all subsequent commands. Re-run login when it expires; scheduler0 logout removes it.

Your signed-in identity is used for createdBy/modifiedBy/deletedBy/archivedBy; there are no --created-by flags.

Point at another environment with --app-url and --base-url:

scheduler0 login --app-url https://app.staging.scheduler0.com --base-url https://api.staging.scheduler0.com

Headless / SSH

The default flow receives the authorization code on a loopback URL (http://127.0.0.1:<port>/callback), which cannot work when the browser runs on a different machine. In an SSH session without a display the CLI switches automatically to a device-code flow: it prints a one-time code and a URL (https://app.scheduler0.com/cli/device); open the URL on any device, enter the code, and the CLI polls until approved. Force it with --device (alias --no-browser).

Authenticating in CI

login is interactive, so in CI set environment variables instead. When SCHEDULER0_API_KEY, SCHEDULER0_API_SECRET and SCHEDULER0_ACCOUNT_ID are present they take priority over any on-disk session.

Variable	Required	Purpose
`SCHEDULER0_API_KEY`	yes	Credential API key
`SCHEDULER0_API_SECRET`	yes	Credential secret (`plaintextSecret` shown once at creation)
`SCHEDULER0_ACCOUNT_ID`	yes	Account the credential belongs to
`SCHEDULER0_BASE_URL`	no	Defaults to `https://api.scheduler0.com`
`SCHEDULER0_ACTOR`	for write commands	Value recorded as `createdBy`/`modifiedBy`/`deletedBy`/`archivedBy`, e.g. `github-actions`
`SCHEDULER0_EXPIRES_AT`	no	RFC3339 expiry for an early local check; otherwise the server rejects an expired credential with `401`

Mint the credential once, interactively:

scheduler0 credentials create --scopes read,write,execute
# save apiKey → SCHEDULER0_API_KEY, plaintextSecret → SCHEDULER0_API_SECRET

Credentials expire after 90 days by default; create a new one and credentials archive the old one before then.

Global flags

--base-url and --account-id are accepted by every command and override the stored session for that invocation.

Commands

Output is JSON unless noted. <id> arguments are positional.

Projects

scheduler0 projects list [--limit 10] [--offset 0] [--order-by date_created] [--order-direction desc]
scheduler0 projects get <project-id>
scheduler0 projects create --name "Billing" --description "Invoice jobs"   # both required
scheduler0 projects update <project-id> --description "New description"    # name cannot be changed
scheduler0 projects delete <project-id>

Jobs

scheduler0 jobs list [--project-id <id>] [--limit 10] [--offset 0] [--order-by date_created] [--order-direction desc]
scheduler0 jobs get <job-id>

scheduler0 jobs create \
  --project-id 123 \
  --executor-id 456 \
  --timezone "UTC" \
  --spec "0 30 * * * *" \
  --data '{"key": "value"}' \
  [--start-date <RFC3339>] [--end-date <RFC3339>] [--retry-max 3] [--status active]

scheduler0 jobs update <job-id> --status inactive     # pause
scheduler0 jobs update <job-id> --status active       # resume
scheduler0 jobs delete <job-id>

--spec is a six-field cron expression (seconds first; see Jobs). For a one-time job omit --spec and pass --start-date. jobs create returns the async task id; check it with scheduler0 async-tasks get <request-id>.

Executors

scheduler0 executors list [--limit 10] [--offset 0] [--order-by date_created] [--order-direction desc]
scheduler0 executors get <executor-id>

scheduler0 executors create \
  --name "orders-webhook" \
  --type webhook_url \
  --webhook-url "https://example.com/hooks/orders" \
  --webhook-method POST \
  --webhook-secret "..." \
  [--description "Sends order emails"] [--tags email,orders] [--payload-aggregation]

scheduler0 executors create \
  --name "reports-lambda" \
  --type cloud_function \
  --cloud-provider aws \
  --region us-east-1 \
  --cloud-resource-url "arn:aws:lambda:us-east-1:123456789012:function:reports" \
  --cloud-api-key "AKIA..." \
  --cloud-api-secret "..."

scheduler0 executors update <executor-id> --name "new-name" --type webhook_url   # name and type are required on update
scheduler0 executors delete <executor-id>
scheduler0 executors test-invoke <executor-id> [--age 24h] [--execution-time <RFC3339>]

--type accepts webhook_url or cloud_function; local executors are created with local-executor register below. --description and --tags are what scheduler0 schedule uses to pick an executor.

Local executor

Runs jobs as a shell command on a machine you control. The CLI pulls jobs assigned to the executor, runs the command, and reports results back. Details of the protocol are in Executors → Local executors.

# 1. Sign in
scheduler0 login

# 2. Register (creates the executor on the server and saves it to ~/.scheduler0/config.json)
scheduler0 local-executor register \
  --name "data-processor" \
  --command "/home/deploy/scripts/process.sh" \
  [--working-dir /home/deploy]

# 3. Assign jobs to the executor id printed above (dashboard, API or `jobs create --executor-id`)

# 4. Run the worker in the foreground
scheduler0 local-executor start [--poll-interval 1m]

Per job, the command receives the job's data on stdin and these environment variables:

Variable	Value
`SCHEDULER0_JOB_ID`	Job id
`SCHEDULER0_JOB_DATA`	Raw `data` field
`SCHEDULER0_JOB_SPEC`	Cron spec
`SCHEDULER0_EXEC_UNIQUE_ID`	Unique id of this execution

A non-zero exit status marks the execution failed. Results are cached in ~/.scheduler0/local-executor.db and reported in batches: every 5 minutes, or immediately once 50 unreported results accumulate. If the server is unreachable, cached jobs keep running and results are flushed when connectivity returns. Ctrl+C stops the worker gracefully.

Executions

scheduler0 executions [--start-date <RFC3339>] [--end-date <RFC3339>] [--project-id <id>] [--job-id <id>] [--limit 10] [--offset 0]
scheduler0 executions analytics --start-date 2026-01-01 --start-time 12:00      # UTC; 12-hour window, hourly buckets
scheduler0 executions totals
scheduler0 executions cleanup-old-logs --account-id <id> --retention-months 6

Credentials

scheduler0 credentials list [--limit 10] [--offset 0] [--order-by date_created] [--order-direction desc] [--output table]
scheduler0 credentials get <credential-id>
scheduler0 credentials create [--scopes read,write,execute]   # prints apiKey and plaintextSecret once
scheduler0 credentials update <credential-id> [--archived]    # PUT: only the archived flag can change; omitting it un-archives
scheduler0 credentials archive <credential-id>
scheduler0 credentials delete <credential-id>

--scopes defaults to read,write,execute; admin can only be granted by an admin-scoped session. apiKey, apiSecret, scopes and expiresAt cannot be changed after creation; to replace a key, credentials create then credentials archive the old one.

AI

# Propose job configurations (POST /ai/prompt)
scheduler0 prompt --prompt "Send weekly reports every Monday at 9 AM" \
  [--purposes reporting] [--events weekly_cycle] [--recipients team@example.com] [--channels email] \
  [--timezone America/New_York] [--locale en-US]

# Intent guardrail only (POST /ai/prompt/classify); English-only --locale
scheduler0 prompt classify --prompt "What is Kubernetes?"

# Full pipeline: create project, match executor, create jobs (POST /ai/schedule)
scheduler0 schedule --prompt "Remind the sales team every Monday at 9am" \
  [--channels email] [--purposes sales] [--timezone America/New_York] [--locale en] \
  [--project-id <id>] [--executor-id <id>]

# Conversation and send-time suggestions; body from --file or stdin
scheduler0 suggestions analyze --file conversation.json
scheduler0 suggestions time --file send-time-request.json

# Prompt request log (GET /ai/prompt-requests)
scheduler0 ai-prompt-requests [--provider openai] [--model ...] [--status success] [--search ...] \
  [--start-date <RFC3339>] [--end-date <RFC3339>] [--order ASC|DESC] [--limit 25] [--offset 0]

# Model catalog and BYOK settings
scheduler0 ai-models                     # GET /ai/models (also: scheduler0 ai-settings models)
scheduler0 ai-settings get               # keys redacted
scheduler0 ai-settings upsert \
  --active-models '[{"provider":"openai","model":"gpt-4.1-mini"},{"provider":"anthropic","model":"claude-sonnet-4-5"}]' \
  --openai-api-key "sk-..." --anthropic-api-key "sk-ant-..." \
  [--openrouter-api-key ...] [--bedrock-access-key-id ...] [--bedrock-secret-key ...] [--bedrock-region ...]

Prompt limits (160 characters; arrays of at most 5 items × 36 characters), quotas, credits and response shapes are described in AI Job Scheduling. schedule returns 409 when no executor can be matched.

Async tasks, features, health

scheduler0 async-tasks get <request-id>   # blocks until the task finishes
scheduler0 features                       # GET /features
scheduler0 healthcheck                    # no auth required

Operator commands (self-hosted)

These call endpoints that require the admin scope; sign in with an admin credential or use SCHEDULER0_* variables for one.

# Accounts
scheduler0 accounts create --name "Team A"
scheduler0 accounts get <account-id>
scheduler0 accounts update <account-id> --name "Team B"
scheduler0 accounts execution-count get <account-id>
scheduler0 accounts execution-count increase <account-id> --count 1000
scheduler0 accounts ai-usage <account-id>
scheduler0 accounts feature add <account-id> --feature-id 3
scheduler0 accounts feature remove <account-id> --feature-id 3
scheduler0 accounts features-all add <account-id>
scheduler0 accounts features-all remove <account-id>
scheduler0 accounts tokens get <account-id>
scheduler0 accounts tokens add <account-id> --amount 500

# Server secret-key rotation (see Account Management)
scheduler0 accounts generate-secret-key [--count 1]          # offline; prints random AES-256 keys
scheduler0 accounts rotate-secret --old-secret-key <OLD_KEY>  # POST /account/rotate-secret

# Backup and restore
scheduler0 backup start
scheduler0 backup restore <file-name>

# Cluster
scheduler0 cluster list-nodes
scheduler0 cluster add-node --node-id 3 --node-address host:7071 --client-address host:9091
scheduler0 cluster remove-node --node-id 3
scheduler0 cluster promote-node --node-id 3
scheduler0 cluster demote-node --node-id 3
scheduler0 cluster transfer-leadership
scheduler0 cluster add-self | remove-self | reset-raft
scheduler0 cluster force-rebuild --seed-node-id 1
scheduler0 cluster dump | schedule-queue | job-executions-cache | job-queues | job-queue-versions

See Account Management, Backup and Restore and Cluster Management for what each endpoint does.

Configuration file

~/.scheduler0/config.json, written by login and local-executor register:

{
  "base_url": "https://api.scheduler0.com",
  "app_url": "https://app.scheduler0.com",
  "api_key": "…",
  "api_secret": "…",
  "account_id": "…",
  "clerk_user_id": "user_…",
  "expires_at": "2026-07-07T12:00:00Z",
  "scopes": ["read", "write", "execute"],
  "local_executor": {
    "id": "42",
    "name": "data-processor",
    "command": "/home/deploy/scripts/process.sh",
    "working_dir": "/home/deploy"
  }
}

local_executor is present only after local-executor register.

Installation​

Install script (macOS, Linux)​

Release binaries​

Docker​

From source​

Sign in​

Headless / SSH​

Authenticating in CI​

Global flags​

Commands​

Projects​

Jobs​

Executors​

Local executor​

Executions​

Credentials​

AI​

Async tasks, features, health​

Operator commands (self-hosted)​

Configuration file​