Scheduling infrastructure for agents.
One schedule for all your agents: Set it. Watch it run. Get proof it worked.
Get API key free → | View docs
# Your agent. On time. Every time. curl -X POST https://api.cueapi.ai/v1/cues \ -H "Authorization: Bearer cue_sk_..." \ -H "Content-Type: application/json" \ -d '{"name":"morning-briefing", "schedule":{"type":"recurring","cron":"0 9 * * *"}, "callback":{"url":"https://your.agent/run"}}' # HTTP 200 { "id": "cue_a3f9b2...", "status": "active" }
POST to /v1/cues. Give it a name, a schedule (cron or ISO timestamp), and the URL to call. Add any payload you want returned. Done.
{ "name": "morning-briefing", "schedule": "0 9 * * *", "callback_url": "https://your.agent/run", "timezone": "America/New_York", "payload": { "task": "check_pipeline" } }
At the scheduled moment, CueAPI calls your webhook with a signed payload. 3 retries with exponential backoff. HMAC-SHA256 verified. At-least-once delivery guaranteed.
{ "execution_id": "exec_8f2a1b", "cue_id": "cue_a3f9b2", "name": "morning-briefing", "scheduled_for": "2026-03-12T09:00:00Z", "attempt": 1, "payload": { "task": "check_pipeline" } }
Run your logic. Report the outcome back. CueAPI logs success, failure, retrying. So you always know if your agents are actually doing the work.
POST /v1/executions/{id}/outcome { "status": "success", "message": "Briefing sent to 3 users", "duration_ms": 1240 }
Your agent said it would handle it. Did it? CueAPI shows every execution with timestamp, cue name, status, attempt number, and duration.
847 jobs. 23 missed in 12 months. That data is real.
Full execution history per cue. Delivery failure vs handler failure, logged separately.
Cron was built for servers in the 1970s. No retries. No history. No visibility. CueAPI is built for agents in 2026.
Every webhook delivery is signed. SSRF protection on all callback URLs.
Exponential backoff. Stale execution recovery. You see exactly what happened.
Claude, ChatGPT, CrewAI, LangChain, custom code. Same five endpoints. One API key.
"The real gap in the market is visibility. Developers don't know their crons are broken until something goes wrong in production." — Socrates, AI Agent, OpenClaw System. First production user, CueAPI closed beta.
"The real gap in the market is visibility. Developers don't know their crons are broken until something goes wrong in production."
— Socrates, AI Agent, OpenClaw System. First production user, CueAPI closed beta.
If you've read the docs, you've learned everything.
# Create a cue curl -X POST https://api.cueapi.ai/v1/cues \ -H "Authorization: Bearer cue_sk_xxx" \ -H "Content-Type: application/json" \ -d '{ "name": "morning-analytics", "schedule": {"type": "recurring", "cron": "0 9 * * *", "timezone": "America/Los_Angeles"}, "callback": {"url": "https://my-agent.com/webhook"}, "payload": {"task": "pull-dashboard", "context_mode": "live"} }' # Response { "id": "cue_a1b2c3d4e5f6", "name": "morning-analytics", "status": "active", "next_run": "2026-03-13T09:00:00-07:00" }
import httpx client = httpx.Client( base_url="https://api.cueapi.ai/v1", headers={"Authorization": "Bearer cue_sk_xxx"} ) cue = client.post("/cues", json={ "name": "morning-analytics", "schedule": {"type": "recurring", "cron": "0 9 * * *"}, "callback": {"url": "https://my-agent.com/webhook"}, "payload": {"task": "pull-dashboard"} }).json() print(cue["next_run"]) # 2026-03-13T09:00:00Z
const res = await fetch("https://api.cueapi.ai/v1/cues", { method: "POST", headers: { "Authorization": "Bearer cue_sk_xxx", "Content-Type": "application/json" }, body: JSON.stringify({ name: "morning-analytics", schedule: { type: "recurring", cron: "0 9 * * *" }, callback: { url: "https://my-agent.com/webhook" }, payload: { task: "pull-dashboard" } }) }); const cue = await res.json(); console.log(cue.next_run); // 2026-03-13T09:00:00Z
$ cueapi create \ --name "morning-analytics" \ --cron "0 9 * * *" \ --url https://my-agent.com/webhook \ --payload '{"task": "pull-dashboard"}' # ✓ Created cue_a1b2c3d4e5f6 # Next run: 2026-03-13 09:00:00 PDT
Same features across all tiers. More volume as you grow.
Save 20% with annual billing. Pro annual: $95.88/year. Scale annual: $470.40/year.
Install → login → first cue in under 30 seconds. Zero friction.
$ pip install cueapi Successfully installed cueapi-0.1.0 $ cueapi login Opening browser to authenticate... ✓ Authenticated as you@email.com $ cueapi quickstart Creating test cue... ✓ Waiting for delivery... ✓ Delivered in 3.8s Your agents now have a clock.
Available CLI commands: cueapi login, cueapi quickstart, cueapi create, cueapi list, cueapi get, cueapi pause, cueapi resume, cueapi delete, cueapi usage, cueapi upgrade.
cueapi login
cueapi quickstart
cueapi create
cueapi list
cueapi get
cueapi pause
cueapi resume
cueapi delete
cueapi usage
cueapi upgrade
CueAPI is tracking NIST AI Agent Standards Initiative and AIUC-1 certification for future compliance. Full security details at cueapi.ai/security.
No. Sign up with just an email. 10 active cues, 300 executions per month, 60 API requests per minute. Free forever.
Per delivery attempt, not per cue fire. If a webhook fails and CueAPI retries 3 times, that counts as 3 executions. This is documented in your usage dashboard.
CueAPI gives you a 24-hour grace period. Cues keep firing during grace. After 24 hours, new executions are blocked until the next billing cycle or you upgrade.
Yes. Upgrades take effect immediately with full access to new limits. Downgrades take effect at the end of your billing period. Excess cues are paused, not deleted.
Yes. Pro annual is $95.88/year (save ~20%). Scale annual is $470.40/year (save ~20%).
Yes. Free, Pro, and Scale only differ in volume: active cues, monthly executions, and request rate. Every feature is available on every plan.
Yes. Install cueapi-worker on your local machine. It pulls scheduled work from CueAPI and runs your scripts locally. No public URL, no webhook server, no tunnel.
CueAPI retries automatically. 3 attempts with exponential backoff (1, 5, 15 minutes by default, configurable up to 10 retries). If all fail, you receive an email notification.
API keys are SHA-256 hashed. Webhooks are signed with HMAC-SHA256 using your per-user secret. All traffic is HTTPS. Callback URLs are validated against SSRF. Full details at cueapi.ai/security.
CueAPI checks for due cues every 5 seconds. Typical delivery latency is 2–3 seconds after the scheduled time. Not suitable for sub-second precision.
Yes. Cancel through the Stripe billing portal. You keep access until the end of your billing period. No refunds for partial periods.
Not yet. If Scale limits are not enough, contact support@vector.build. Custom plans are on the roadmap.
All major credit and debit cards via Stripe.
CueAPI targets high availability but does not offer a formal SLA yet. Current status is always visible at status.cueapi.ai. Failures delay delivery but never lose schedules.
Let the agent be the brain. Let CueAPI be the clock.
Get API key free → | Read the docs
Agent API Reference · About · Pricing · Blog · Security · Privacy · Terms · Docs · Dashboard · X / Twitter · Status
© Vector Apps Inc 2026
Tutorials, guides, and updates from the CueAPI team on AI agent scheduling, cron replacement, and webhook infrastructure.
Tutorial · March 14, 2026 · 5 min read
Create your first scheduled cue with CueAPI. Three methods: dashboard, CLI, or API.
Blog · March 14, 2026 · 4 min read
Silent cron failures are the hardest bug in agent systems.
Blog
The AI agent gold rush has produced countless demos showcasing impressive capabilities. Yet most crumble in production. Reliability trumps capability when it comes to systems people actually depend on.
RSS Feed: Subscribe via RSS
← Back to cueapi.ai
REST API for scheduling agent tasks. 24 user-facing endpoints. HMAC-SHA256 signed webhooks with v1= signature prefix. Per-user secrets. At-least-once delivery. Webhook + Worker transport. Free tier: 10 cues, 300 executions/month.
Base URL: https://api.cueapi.ai · Auth: Bearer cue_sk_{32 hex chars} · Docs: https://api.cueapi.ai/docs
REGISTER FIRE RUN REPORT POST /v1/cues --> CueAPI fires --> Handler runs --> POST /outcome on schedule your code success/error
Two transport modes: Webhook (CueAPI POSTs to your URL) or Worker (local daemon polls for jobs via pip install cueapi-worker — no public URL needed). Most agents use Worker.
Every webhook delivery includes X-CueAPI-Signature: v1=a1b2c3d4e5f6... header. Verify with HMAC-SHA256 using your per-user secret.
# Python import hmac, hashlib expected = "v1=" + hmac.new(secret.encode(), body, hashlib.sha256).hexdigest() // Node.js const expected = "v1=" + crypto.createHmac("sha256", secret).update(body).digest("hex");
When all retries are exhausted, CueAPI can notify you automatically. Set on_failure when creating a cue:
{ "on_failure": { "email": true, "webhook": "https://my-agent.com/alert", "pause": false } }
All three can be used together. Default is email only.
Default: 3 retries, exponential backoff (1 min, 5 min, 15 min). Configurable up to 10 retries. run_count = successful deliveries. fired_count = total attempts including retries.
All plans include webhook + worker transport, retry config, on_failure escalation, and full execution history.
If a worker loses auth (401 errors), recover without redeploying:
cueapi-worker login # Re-authenticate via email magic link cueapi-worker regenerate-key # Generate new API key (revokes old)
The worker auto-detects 401 and attempts recovery with 3 retries and 30s backoff.
pip install cueapi # CLI pip install cueapi-worker # Local worker daemon # Register curl -X POST https://api.cueapi.ai/v1/auth/register \ -H "Content-Type: application/json" \ -d '{"email": "agent@example.com"}' # Create a cue curl -X POST https://api.cueapi.ai/v1/cues \ -H "Authorization: Bearer cue_sk_..." \ -H "Content-Type: application/json" \ -d '{"name":"test","schedule":{"type":"once","at":"2026-03-13T10:01:00Z"},"callback":{"url":"https://my-endpoint.com/hook"}}'
No credit card required for free tier. Get API key free →
See full API reference →
Cron runs jobs silently. CueAPI runs jobs with retries, failure detection, execution proof, timezone handling, and observability.
Implementation details at Agent API Reference →
Temporal is a workflow orchestration platform for engineering teams. CueAPI is a scheduling API for solo builders. Do you need orchestration or scheduling?
Inngest is event-driven for serverless. CueAPI is schedule-driven for agents. Local worker support, no public URL needed.
Schedule agent turns with LLMs, automate monitoring checks, run content pipelines on time. CueAPI handles timing. Your agent handles the work.
Every schedule matters.
A report that runs every morning. A check that fires every hour. A follow-up sent three days later. A backup triggered at midnight. A tweet posted at the right moment.
Right now, all of it runs on cron jobs and hope. Something fails, nobody knows. Something misses, nobody notices. The agent says it handled it. Maybe it did. Maybe it didn't.
CueAPI makes scheduling visible. Set a schedule through the API. CueAPI fires on time, retries on failure, and logs every execution with proof of what happened.
Every job becomes something you can verify.
CueAPI is owned by Vector, co-founded by Mike Molinet and Govind Kavaturi.
Last updated: March 13, 2026 · Operated by Vector Apps Inc.
This Privacy Policy describes how Vector Apps Inc. ("Vector Apps," "CueAPI," "the Service") collects, uses, and protects information when you use the CueAPI scheduling API and related services.
Account Information: email address. API Usage Data: cue configurations, execution logs, usage metrics, API request logs. Payment Information: handled entirely by Stripe. Webhook Payloads: stored for delivery only. Support Messages: stored indefinitely until resolved.
Operate the scheduling service, deliver webhooks, process payments, provide execution logs, enforce rate limits, monitor system health, send transactional emails, respond to support requests.
CueAPI does not use your data for advertising. CueAPI does not sell your data. CueAPI does not train AI models on your data.
Account data: lifetime of account. Execution logs: 90 days. Cue configurations: until you delete them. API request logs: 30 days.
Access your data via the API. Export cue configurations and execution history. Delete cues and history. Request full account deletion. GDPR rights for EU users.
Railway (hosting), Stripe (payments), Resend (email), Cloudflare (DNS/CDN), UptimeRobot (monitoring).
Email: support@vector.build · Company: Vector Apps Inc. · Location: Palo Alto, California, United States
These Terms of Service govern your use of CueAPI, a scheduling API operated by Vector Apps Inc.
CueAPI provides a scheduling API for creating scheduled tasks ("cues"), receiving webhook deliveries or pulling executions via a local worker, and reporting execution outcomes.
Free tier: 10 active cues, 300 executions/month, 60 req/min. No credit card. Pro: $9.99/mo or $95.88/year. Scale: $49/mo or $470.40/year. Upgrades take effect immediately. Downgrades at end of billing period.
Use CueAPI for any lawful purpose. Do not send webhooks to URLs you don't own, access other users' data, circumvent rate limits, conduct DoS attacks, store illegal content, or reverse-engineer the Service.
At-least-once delivery. Your handler may receive the same execution more than once. Use execution_id as your dedup key. CueAPI targets high availability but does not guarantee specific uptime percentages.
API keys are hashed with SHA-256. Webhook deliveries signed with HMAC-SHA256. All API communication requires HTTPS. Full privacy practices in the Privacy Policy.
CueAPI is provided "as is" without warranties. Total liability limited to amount paid in preceding 12 months.
Close your account by contacting support@vector.build. Data retained 30 days for recovery, then permanently deleted.
Support requests submitted via the API (POST /v1/support) are stored in our systems and tracked until resolved. Stored data includes your email, plan, message, severity, and ticket type. See our Privacy Policy for full details on data retention and your rights.
Governed by the laws of the State of California.
Every API request requires a Bearer token (cue_sk_{32 hex chars}). API keys are SHA-256 hashed before storage. Never stored in plaintext. Auth cached in Redis with 5-minute TTL, falls back to PostgreSQL if Redis is unavailable.
Every delivery signed with HMAC-SHA256 using your per-user webhook secret (whsec_). Signing input: {timestamp}.{json_payload}. Verify with constant-time comparison. Reject deliveries older than 5 minutes.
X-CueAPI-Signature (HMAC-SHA256), X-CueAPI-Timestamp, X-CueAPI-Cue-Id, X-CueAPI-Execution-Id, X-CueAPI-Scheduled-For, X-CueAPI-Attempt, Content-Type: application/json, User-Agent: CueAPI/1.0.
HTTPS only (TLS 1.2+). SSRF protection: blocks localhost, private networks, link-local, cloud metadata endpoints, IPv6 equivalents. DNS resolution validated.
Free: 60/min. Pro: 200/min. Scale: 500/min. Sliding window algorithm. HTTP 429 with Retry-After header on violation.
At-least-once delivery. 3 retries with exponential backoff (configurable 1-10). Cues checked every 5 seconds. 30-second webhook timeout. Stale execution recovery. Leader election via Redis. PostgreSQL persistence.
POST /v1/support protected by 5 gates: authentication, rate limiting (5/hour), account age (>1 hour), product usage (≥1 cue), content validation (min 10 char subject, 30 char message).
Railway (hosting), Stripe (payments), Resend (email), Cloudflare (DNS/CDN), UptimeRobot (monitoring). CueAPI does not share payloads, execution data, or API keys with subprocessors.
Report vulnerabilities to support@vector.build. CueAPI will respond within 48 hours.
Current system status: status.cueapi.ai