Fliq

Reliable HTTP job scheduling for developers and engineering teams.

Stop reinventing cron. Fliq lets you schedule one-off HTTP calls and recurring cron jobs with a single API call — with built-in retries, crash recovery, and per-execution history. Whether you're triggering a webhook at a specific time or running a recurring data pipeline, Fliq handles the reliability layer so you don't have to.

---

Why Fliq

For startups — Ship faster. Don't burn a sprint building a job queue. POST a job, move on. Fliq handles delivery, retries, and failure visibility out of the box.

For developers — No SDKs, no magic. A clean REST API, API token auth (fliq_sk_*), idempotency keys, and full execution history per job. Works with any HTTP endpoint. Delivery is at-least-once — every outbound request carries a stable X-Fliq-Delivery-Id header so your endpoint can dedupe.

For engineering leaders — One less piece of infrastructure to own. Postgres-native, no Redis or Kafka required. Sub-2s pickup latency. At-least-once execution with crash-safe retries via row-level locking. Designed to scale.

---

Pricing

Plan	Executions	Price
Free	5,000 / day	$0
Pay-as-you-go	Unlimited	$1 per 100,000 executions

No seat fees. No contracts. Pay only for what you execute.

---

What's built

Area	Status
One-off job scheduling — create, claim, execute, retry	✅ Done
Cron schedules — recurring jobs with pause/resume	✅ Done
At-least-once execution (FOR UPDATE SKIP LOCKED + reaper)	✅ Done
Crash recovery (heartbeat + reaper process)	✅ Done
API token auth (fliq_sk_*) + Clerk JWT	✅ Done
Per-user job isolation (ownership enforced at query level)	✅ Done
Credit system — free tier + pay-as-you-go via Stripe	✅ Done
CI pipeline (lint, tests, migrations against real Postgres)	✅ Done

---

System map

[ Client ]
    │  REST API
    ▼
[ server ]  ──────────────────────────────────┐
    │                                         │
    ▼                                         ▼
[ PostgreSQL ] ◄──────────────── [ scheduler ]
                                  Worker + Reaper + Dispatcher + Executor

---

Stack

Concern	Choice
Language	Go 1.25
Web framework	Gin
Database	PostgreSQL via pgx/v5
Migrations	goose
Auth	Clerk (JWT RS256 via JWKS) + API tokens (HS256 fallback for local dev)
Billing	Stripe — webhooks + credit system
Config	caarlos0/env — struct tags, no .env files in Go code
Linter	golangci-lint v2

---

Local dev

# Prerequisites: Docker, direnv, goose
eval "$(direnv hook zsh)"   # if not already in ~/.zshrc

docker compose up -d postgres
direnv allow
goose -dir ./migrations postgres "$DATABASE_URL" up

go run ./cmd/server        # terminal 1
go run ./cmd/scheduler     # terminal 2

See CLAUDE.md for the full local setup guide and coding conventions.

---

Roadmap

Phase 1 — Core backend ✅

• Job CRUD, worker, reaper, retry with backoff
• Cron schedules with pause/resume
• At-least-once execution with crash-safe retries via Postgres row-level locking
• API token auth + Clerk JWT; jobs scoped to authenticated users
• Credit system: free tier (5k/day) + pay-as-you-go via Stripe
• CI: lint + test + migrations on every PR

Phase 2 — Deployment 🔄 In progress

• Docker images (Dockerfile.server, Dockerfile.scheduler, Dockerfile.migrate)
• Deploy to K8S on rented VM
• Staging and production environments
• Terraform for infra provisioning (Enkidu)

Phase 3 — Observability

• OpenTelemetry instrumentation (traces + metrics)
• Prometheus + Grafana dashboards
• Key metrics: job pickup latency, reaper rescue rate, worker lifetime, API p99

Phase 4 — Frontend

• Dashboard: job list, execution history, schedule management
• Docs integrated into the website — simple, developer-focused