// Use this skill whenever someone needs a backend, or a task touches backend or cloud infrastructure: at minimum read it to check relevance, then stop if the task is not actually backend/cloud work, or use it to provision and manage that backend with the InsForge CLI if it is. Covers projects, SQL, migrations, RLS policies, functions, storage buckets, frontend deployments, compute services, secrets/env vars, AI/OpenRouter key setup, Stripe payment keys/catalog/products/prices/webhooks, schedules, logs, diagnostics, import/export, **declarative auth redirect URLs via `insforge.toml`** (applied with `config apply`), or **managing backend branches** (creating a branch project to test risky schema/auth/RLS changes, merging a branch back to prod, resolving merge conflicts). For app code with @insforge/sdk, use the insforge skill instead.
Use this skill whenever someone needs a backend, or a task touches backend or cloud infrastructure: at minimum read it to check relevance, then stop if the task is not actually backend/cloud work, or use it to provision and manage that backend with the InsForge CLI if it is. Covers projects, SQL, migrations, RLS policies, functions, storage buckets, frontend deployments, compute services, secrets/env vars, AI/OpenRouter key setup, Stripe payment keys/catalog/products/prices/webhooks, schedules, logs, diagnostics, import/export, **declarative auth redirect URLs via `insforge.toml`** (applied with `config apply`), or **managing backend branches** (creating a branch project to test risky schema/auth/RLS changes, merging a branch back to prod, resolving merge conflicts). For app code with @insforge/sdk, use the insforge skill instead.
If not authenticated: npx @insforge/cli login
If no project linked: npx @insforge/cli create (new — for non-interactive agent use, pass --json with --name, --template, --org-id, --region to skip all prompts; -y alone is NOT enough) or npx @insforge/cli link (existing)
Important: In InsForge, API keys are full-access admin keys, equivalent to service role keys on other platforms. Treat them as server-only secrets and never expose them in frontend code or public env vars.
Global Options
Flag
Description
--json
Structured JSON output AND skip value-collection prompts (text / select). Errors out if any required param is missing. Combine with -y for destructive commands so Y/N confirmations are skipped too.
-y, --yes
Auto-accept Y/N confirmation prompts (delete? overwrite? etc.). Does NOT skip value-collection prompts — use --json for that.
All examples below use npx @insforge/cli. Never call insforge directly.
Exit Codes
Code
Meaning
0
Success
1
General error (e.g., HTTP 400+ from function invoke)
2
Not authenticated
3
Project not linked
4
Resource not found
5
Permission denied
Environment Variables
Variable
Description
INSFORGE_ACCESS_TOKEN
Override stored access token
INSFORGE_PROJECT_ID
Override linked project ID
INSFORGE_EMAIL
Email for non-interactive login
INSFORGE_PASSWORD
Password for non-interactive login
Commands
Authentication
npx @insforge/cli login — OAuth (browser) or --email for password login. See references/login.md
npx @insforge/cli link — link directory to existing project
npx @insforge/cli current — show current user + linked project
npx @insforge/cli list — list all orgs and projects
npx @insforge/cli metadata — show backend metadata (auth config, database tables, storage buckets, edge functions, AI models, realtime channels). Use --json for structured output. Run this first to discover what's configured before building features.
npx @insforge/cli ai setup — fetch the linked project's active OpenRouter key from InsForge and write OPENROUTER_API_KEY to .env.local for local server-side development. Use --env-file <path> for non-standard projects.
Run this before implementing new OpenRouter/OpenAI SDK app code. Keep OPENROUTER_API_KEY server-only: never rename it to NEXT_PUBLIC_*, VITE_*, PUBLIC_*, or REACT_APP_*. For production deploys, add the same OPENROUTER_API_KEY to the hosting provider's server/runtime environment.
Payments — npx @insforge/cli payments
npx @insforge/cli payments status — show Stripe key, account, sync, and webhook status
npx @insforge/cli payments config / config set / config remove — manage Stripe test/live secret keys. See references/payments.md
npx @insforge/cli payments sync [--environment test|live|all] — sync products, prices, customers, and subscriptions from Stripe
npx @insforge/cli payments webhooks configure <environment> — create or recreate the managed Stripe webhook endpoint
npx @insforge/cli payments catalog --environment <env> — inspect mirrored products and prices together
npx @insforge/cli payments history --environment <env> — admin/debug payment history reads
⚠️ Private preview. Payments are a new feature; older backends may not expose /api/payments.
Availability: If the CLI says Payments are not available on this backend, stop and ask the developer/admin to enable payments or upgrade the self-hosted InsForge instance. Do not work around this by storing Stripe secret keys with generic secrets or embedding Stripe secret keys in app code.
Agents should default to --environment test while building. Only use live after the developer explicitly approves production Stripe changes.
Deploy a frontend application (static site / SPA / Next.js / etc.) to Vercel,
managed through InsForge. For backend container workloads see Backend Compute
Services below.
npx @insforge/cli deployments deploy [dir] — deploy frontend app from its source directory. See references/deployments-deploy.md
npx @insforge/cli deployments list — list deployments
npx @insforge/cli deployments status <id> [--sync] — get deployment status (--sync fetches from Vercel)
Deploy and manage backend containerized services (APIs, workers, microservices).
Each service runs as a Docker container reachable via a public endpoint — https://<name>-<project>.fly.dev for HTTP services (the default), or <name>-<project>.fly.dev:<port> raw TCP for services deployed with --protocol tcp (Redis, Postgres-protocol, etc.). See TCP services below.
For frontend hosting see Frontend Deployments above.
🔧 Implementation note (for agents): InsForge runs compute on Fly.io
under the hood, but DO NOT use flyctl directly to deploy or manage
these services. The Fly account, org, IP allocation, and machine ownership
all live on the InsForge cloud — flyctl invoked with the user's own credentials
will land in the wrong org and fail with unauthorized. Always use
npx @insforge/cli compute …. The CLI is just an HTTP client that calls the
InsForge backend; the backend talks to Fly. No flyctl and no Fly token
are needed locally.
⚠️ In progress. Compute services are still in development; the API and CLI may change.
Availability: Compute requires the backend to have Fly.io configured. If not enabled, the API returns COMPUTE_SERVICE_NOT_CONFIGURED with setup instructions in nextActions. Follow those instructions.
npx @insforge/cli compute list — list all compute services (name, status, image, CPU, memory, endpoint)
npx @insforge/cli compute get <id> — get service details
npx @insforge/cli compute deploy [dir] --name <name> [--port] [--cpu] [--memory] [--region] [--env <json> | --env-file <path>] — source mode: requires flyctl on PATH; no local Docker daemon needed. CLI shells out to flyctl deploy --remote-only --build-only using a short-lived per-app deploy token minted by InsForge cloud (the user never sees a Fly token). Build runs on Fly's remote builder; image is pushed to registry.fly.io; cloud launches the machine.
npx @insforge/cli compute deploy --image <url> --name <name> [--port] [--cpu] [--memory] [--region] [--protocol <http|tcp>] [--env <json> | --env-file <path>] — image mode: deploys a pre-built image from any registry. Nothing needed locally beyond the InsForge CLI. Best for CI/CD pipelines and off-the-shelf images like nginx:alpine. Prefer --env-file <path> over inline --env <json> for >1 secret. Pass --protocol tcp for raw-TCP services like Redis (see TCP services below).
npx @insforge/cli compute update <id> [--image] [--port] [--cpu] [--memory] [--region] [--env <json> | --env-set KEY=VALUE | --env-unset KEY] — update service config. --env-set/--env-unset are repeatable and merge with existing env — use these to rotate one secret without restating the rest. --env <json> replaces wholesale and is mutually exclusive with the merge flags.
npx @insforge/cli compute stop <id> — stop a running service
npx @insforge/cli compute start <id> — start a stopped service
npx @insforge/cli compute events <id> [--limit 50] — Fly machine lifecycle events (start/stop/exit/restart). Container stdout/stderr is NOT surfaced in v1 — that's roadmap work and will reuse the freed-up compute logs command name when it lands. To debug a crash-looping container today, reproduce locally with the same image.
npx @insforge/cli compute delete <id> — destroy the service and its Fly.io resources. Permanent. Audit log captures the full config (incl. encrypted env blob) on delete for reconstruction. Dashboard adds a type-to-confirm gate; the CLI does not — guard scripted deletes carefully.
Secrets — npx @insforge/cli secrets
npx @insforge/cli secrets list [--all] — list secrets (values hidden; --all includes deleted)
npx @insforge/cli secrets get <key> — get decrypted value
Manage auth.allowed_redirect_urls declaratively via insforge.toml (project root). For changing redirect URLs, prefer this over PUT /api/auth/config — the CLI gates on backend version and surfaces a clean skip if the backend predates the field, where the raw PUT may 200-and-silently-drop.
npx @insforge/cli config export [--out insforge.toml] [--force] — pull live config into TOML. Sections the backend doesn't expose are omitted.
npx @insforge/cli config plan [--file insforge.toml] — diff TOML vs. live state; shows which changes will apply vs. be skipped on the connected backend.
npx @insforge/cli config apply [--file insforge.toml] [--dry-run] [--auto-approve] — apply the TOML. Per-change capability gate: supported changes apply; unsupported go to skipped[] with an upgrade message and no PUT is issued for them. --json returns { plan, applied[], skipped[] }.
If apply returns skipped: [...], surface verbatim. The user's backend predates this section. Tell them which sections were skipped and to upgrade; do not retry, do not bypass with curl (silent drop). Sample message: "I tried to set auth.allowed_redirect_urls but your backend is on an older version that doesn't support it yet. Upgrade your backend and re-run npx @insforge/cli config apply."
TOML is for knobs only — never embed programs. SQL → db migrations. Function code → functions deploy. Compute → compute deploy. Frontend → deployments deploy. TOML carries booleans, strings, and arrays — anything bigger lives in its own file managed by a dedicated CLI command.
Scope today: only auth.allowed_redirect_urls. Password policy, SMTP, OAuth providers, custom subdomain, and similar will land in TOML later. For other auth.config fields today, use the dashboard.
Test risky schema, RLS, auth, or function changes in an isolated branch project before applying them to prod. A branch shares the parent's JWT_SECRET (so the same users authenticate) but gets a fresh EC2 + database + API_KEY / ANON_KEY. See references/branch.md for the decision guide and lifecycle command details.
⚠️ Backend version requirement. Branching is still in development and only available on InsForge backend 2.1.0 or later. On older backends, npx @insforge/cli branch commands will fail. If you hit an error, contact the InsForge team.
Compute (and optionally apply) the 3-way merge to parent. Conflict path exits with code 2. See branch-merge.
npx @insforge/cli branch reset <name> [-y]
Wipe all changes on the branch and restore its database to T0 (parent's snapshot at branch creation). Same EC2 / appkey / API_KEY — only DB content rewinds. Works from readyormerged. See branch-reset.
npx @insforge/cli branch delete <name> [-y]
Delete a branch and reclaim its EC2. Auto-switches back to parent if currently on the deleted branch.
Typical flow:
npx @insforge/cli branch create feat-rls --mode schema-only
# context now points at the branch — re-source .env if your dev server caches it# branch runtime starts empty — redeploy any functions/website/compute you'll exercise# ... apply your schema / RLS / auth changes via db migrations / SQL ...
npx @insforge/cli branch merge feat-rls --dry-run --save-sql /tmp/diff.sql
# review /tmp/diff.sql; on conflict CLI exits 2
npx @insforge/cli branch merge feat-rls
# parent now has the changes — redeploy functions / website / compute as needed
npx @insforge/cli branch delete feat-rls
Got into a bad state on the branch?npx @insforge/cli branch reset <name> rewinds the branch's database back to T0 (the parent snapshot at branch creation) without touching the EC2 or API keys — cheaper than delete + recreate and the SDK's INSFORGE_URL / ANON_KEY stay valid. Works from both ready and merged (a merged branch reset re-opens the same slot for another round of changes).
Diagnostics — npx @insforge/cli diagnose
Run with no subcommand for a full health report across all checks.
npx @insforge/cli diagnose — full health report (runs all diagnostics)
npx @insforge/cli diagnose --ai "<issue description>" — hand a natural-language problem description (error, failing URL, HTTP status) to the InsForge debug agent; returns a diagnosis plus suggested solutions
npx @insforge/cli diagnose db [--check <checks>] — database health checks. Checks: connections, slow-queries, bloat, size, index-usage, locks, cache-hit (default: all)
npx @insforge/cli diagnose logs [--source <name>] [--limit <n>] — aggregate error-level logs from all backend sources. Default limit: 100
Analytics Integration — npx @insforge/cli posthog
npx @insforge/cli posthog setup — ensures the InsForge dashboard has a PostHog connection (auto-provisioning / OAuth if missing, no-op if already connected), then prints the official PostHog wizard command (npx -y @posthog/wizard@latest) for the user to run themselves to wire PostHog into the app code. The CLI itself does not spawn the wizard. See references/posthog.md.
For agents: the printed wizard command is interactive — prompts on stdin, opens a browser, won't work via the agent shell or ! prefix. Ask the user to switch to their real terminal app and run it there.
⚠️ Beta. PostHog integration is rolling out across InsForge Cloud; older cloud projects and self-hosted backends may not expose /integrations/posthog/v1/* yet. If the CLI fails with PostHog connect flow unavailable (HTTP 404), the project doesn't have PostHog enabled yet — wait for the rollout or ask the InsForge team to enable it. Do not work around this by pulling a phc_ key from a separate PostHog account and embedding it in the app's env: events will flow to PostHog but the InsForge dashboard's Analytics page reads from a server-side OAuth-backed posthog_connections row that only posthog setup populates — without it the page stays empty even though the integration "looks" wired.
For writing application code with the InsForge SDK, use the insforge (SDK) skill instead, and use the npx @insforge/cli docs <feature> <language> to get specific SDK documentation.
Non-Obvious Behaviors
Functions invoke URL: invoked at {oss_host}/functions/{slug} — NOT /api/functions/{slug}. Exits with code 1 on HTTP 400+.
Secrets delete is soft: marks the secret inactive, not destroyed. Restore with npx @insforge/cli secrets update KEY --active true. Use --all with secrets list to see inactive ones.
Storage delete-bucket is hard: deletes the bucket and every object inside it permanently.
db rpc uses GET or POST: no --data → GET; with --data → POST.
db migrations use timestamped files: migration filenames use YYYYMMDDHHmmss_name.sql, for example 20260418091500_create-posts.sql.
db migrations up supports safe batch modes: npx @insforge/cli db migrations up <filename|version> applies one explicit local target. npx @insforge/cli db migrations up --to <version|filename> and npx @insforge/cli db migrations up --all apply pending files in ascending version order and stop on the first failure.
db migrations run inside a backend-managed transaction: do not put BEGIN, COMMIT, or ROLLBACK in migration files.
The live database schema is the source of truth: before writing a migration, and again if a migration fails, inspect the current database state first (db tables / indexes / policies / triggers / functions, plus db migrations list) and then adjust the migration statements to match reality. Do not assume local files are still current.
Compute deploy has two modes.compute deploy [dir] shells out to flyctl deploy --remote-only --build-only against your source dir using a short-lived per-app token the cloud mints for that one deploy — requires flyctl on PATH but no local Docker daemon (the build runs remotely on Fly's builder). The token is attenuated to one app + builder/wg with else: deny so it cannot reach any other app or org-level endpoint, and it auto-expires after ~20 min. compute deploy --image <url> deploys a pre-built image from any registry — nothing needed locally, best for CI or off-the-shelf images like nginx:alpine. Don't use flyctl outside this CLI flow with your own credentials — the Fly account is InsForge's, you'd 401.
Compute endpoints use .fly.dev: Services get a public URL at https://{name}-{projectId}.fly.dev. Custom domains require DNS configuration.
Schedules accept two cron formats: 5-field cron (minute hour day month day-of-week, e.g. */5 * * * *) or pg_cron interval syntax for sub-minute cadence (e.g. 30 seconds). 6-field cron with seconds (Quartz/Spring's */2 * * * * *) is not supported — use the interval form for sub-minute work. Headers can reference secrets with ${{secrets.KEY_NAME}}.
Payments use Stripe as source of truth: use payments config set for Stripe keys, payments sync before relying on existing catalog data, and create a new Stripe price instead of editing amount/currency. Runtime checkout and customer portal integration belongs in the insforge SDK skill.
config apply is version-aware: per-project backends drift in version. The CLI probes /api/metadata and gates auth.allowed_redirect_urls on whether the backend exposes it; if not, the change lands in skipped[] with an upgrade message and no PUT is issued. Never bypass with curl to "force" a skipped field — older backends may 200-and-silently-drop. Surface skips to the user and ask them to upgrade.
Common Workflows
Configure auth redirect URLs (insforge.toml)
# 1. Pull current config to insforge.toml (writes in project root)
npx @insforge/cli --json config export# 2. Edit insforge.toml — add/remove entries under [auth] allowed_redirect_urls# 3. Preview the diff
npx @insforge/cli --json config plan
# 4. Apply (use --yes for non-interactive — required in --json mode)
npx @insforge/cli --json --yes config apply
If apply returns a non-empty skipped[], the user's backend predates this section — surface the entries verbatim with the upgrade ask. Don't retry, don't fall back to curl (older backends may 200-and-silently-drop). See references/config.md.
Set up database schema with migrations
# Inspect the current live schema first
npx @insforge/cli db tables
npx @insforge/cli db indexes
npx @insforge/cli db policies
npx @insforge/cli db migrations list
# Sync applied remote migration history locally
npx @insforge/cli db migrations fetch
# Create the next schema migration file
npx @insforge/cli db migrations new create-posts
# Edit migrations/20260418091500_create-posts.sql with CREATE TABLE / ALTER TABLE / policies# Apply pending migrations safely
npx @insforge/cli db migrations up --all
db query and migrations run as project_admin. Use migrations for schema changes and RLS policies, including RLS on documented InsForge-managed tables. Use db query for row changes and inspection. In migrations, FK to users with auth.users(id) and use auth.uid() in RLS policies.
Manage database migrations
# Inspect remote migration history
npx @insforge/cli db migrations list
# Sync applied remote migrations into migrations/
npx @insforge/cli db migrations fetch
# Create the next local migration file
npx @insforge/cli db migrations new create-posts
# Apply all pending local migrations
npx @insforge/cli db migrations up --all
Deploy an edge function
# --file is required; point it at your source wherever it lives
npx @insforge/cli functions deploy my-handler --file ./my-handler.ts
npx @insforge/cli functions invoke my-handler --data '{"action": "test"}'
Deploy frontend
Always verify the local build succeeds before deploying. Local builds are faster to debug and don't waste server resources. After the build passes, deploy the project source directory (usually .), not dist/ or other generated build output.
Environment variables are required. Frontend apps need env vars (API URL, anon key) to connect to InsForge at runtime. Deploying without them produces a broken app. Before deploying, you must ensure env vars are set using one of these two approaches:
Option A — Persistent env vars (recommended): Set once, applied to every future deployment automatically. Best for projects that will be redeployed.
# Check what's already set
npx @insforge/cli deployments env list
# Set the vars your app needs (use the correct framework prefix)
npx @insforge/cli deployments envset VITE_INSFORGE_URL https://my-app.us-east.insforge.app
npx @insforge/cli deployments envset VITE_INSFORGE_ANON_KEY ik_xxx
# Deploy the project source — persistent env vars are applied automatically
npx @insforge/cli deployments deploy .
Option B — Inline --env flag: Pass env vars as JSON directly on the deploy command. Useful for one-off deploys or overriding persistent vars.
# 1. Build locally first
npm run build
# 2. Ensure env vars are set (check existing, add missing)
npx @insforge/cli deployments env list
npx @insforge/cli deployments envset VITE_INSFORGE_URL https://my-app.us-east.insforge.app
npx @insforge/cli deployments envset VITE_INSFORGE_ANON_KEY ik_xxx
# 3. Deploy the project source directory
npx @insforge/cli deployments deploy .
Environment variable prefix by framework:
Framework
Prefix
Example
Vite
VITE_
VITE_INSFORGE_URL
Next.js
NEXT_PUBLIC_
NEXT_PUBLIC_INSFORGE_URL
Create React App
REACT_APP_
REACT_APP_INSFORGE_URL
Astro
PUBLIC_
PUBLIC_INSFORGE_URL
SvelteKit
PUBLIC_
PUBLIC_INSFORGE_URL
Pre-deploy checklist:
npm run build succeeds locally
Env vars are set — run deployments env list to verify, or pass --env on the deploy command
All env vars use the correct framework prefix
Deploy the project source directory (usually .), not dist/, build/, or .next/
Edge function directories excluded from frontend build (if applicable)
Never include node_modules, .git, .env, or .insforge in the upload
Framework build output is configured correctly (dist/, build/, .next/, etc.)
Deploy a Docker container (compute service)
Two modes — pick by what you have. Both deploy to the same Fly.io infrastructure.
Source mode (you have a Dockerfile; needs flyctl on PATH but NO local Docker daemon):
# Install flyctl once: curl -L https://fly.io/install.sh | sh# Project layout: Dockerfile + your app code
$ ls
Dockerfile app.py requirements.txt
# One command:
npx @insforge/cli compute deploy . --name my-api --port 8000
# CLI mints a per-app, attenuated Fly deploy token from InsForge cloud# (~20 min TTL, scoped to one app, else: deny blocks org-wide reads), then# shells out to `flyctl deploy --remote-only --build-only` so the build runs# on Fly's remote builder and is pushed to registry.fly.io. Cloud launches# the machine and returns the URL. NO local Docker daemon needed.
Off-the-shelf image (no Docker required):
npx @insforge/cli compute deploy --image nginx:alpine --name my-api --port 80 --region iad
npx @insforge/cli compute list
# Service is running with a public https://{name}-{project}.fly.dev endpoint# No flyctl, no FLY_API_TOKEN, no local Docker required.
Pre-built image you pushed yourself (CI/CD, custom registry — no Docker required locally):
# Built + pushed elsewhere (GitHub Actions, your CI, etc.)
npx @insforge/cli compute deploy --image ghcr.io/you/app:v1 --name my-api --port 8000
TCP services (Redis, Postgres-protocol, etc.):
# Default is HTTP; pass --protocol tcp for raw TCP pass-through at the Fly edge.
npx @insforge/cli compute deploy \
--image redis:7-alpine \
--name cache \
--protocol tcp \
--port 6379
# Endpoint shape is <host>:<port>, e.g. cache-<project>.fly.dev:6379.# Connect with `redis-cli -h <host> -p 6379`.
TCP services are reachable from the public internet. Configure auth in your container (e.g. start Redis with --requirepass <secret> via --env). Storage is ephemeral in v1; AOF/RDB persistence requires volumes (not yet supported by InsForge Compute).
CPU tiers:shared-1x (default), shared-2x, performance-1x, performance-2x, performance-4xMemory options: 256, 512 (default), 1024, 2048, 4096, 8192 MB
Regions:iad (default), sin, lax, lhr, nrt, ams, syd
Source mode requires flyctl on PATH (no Docker). The CLI never asks the user for FLY_API_TOKEN — the cloud mints a short-lived, app-scoped token per deploy (~20 min, else: deny) and passes it through env to the flyctl subprocess. Tokens cannot deploy or read any other app, even within InsForge's Fly org. Image mode (the examples above) needs neither flyctl nor a token.
Backup and restore database
npx @insforge/cli db export --output backup.sql
npx @insforge/cli db import backup.sql
InsForge accepts two cron formats: standard 5-field cron expressions, or pg_cron interval syntax for sub-minute cadence. 6-field cron expressions with seconds (Quartz/Spring style) are NOT supported — use the interval form below for sub-minute work.
5-field cron format:
┌─────────────── minute (0-59)
│ ┌───────────── hour (0-23)
│ │ ┌─────────── day of month (1-31)
│ │ │ ┌───────── month (1-12)
│ │ │ │ ┌─────── day of week (0-6, Sunday=0)
│ │ │ │ │
* * * * *
Expression
Description
* * * * *
Every minute
*/5 * * * *
Every 5 minutes
0 * * * *
Every hour (at minute 0)
0 9 * * *
Daily at 9:00 AM
0 9 * * 1
Every Monday at 9:00 AM
0 0 1 * *
First day of every month at midnight
30 14 * * 1-5
Weekdays at 2:30 PM
Interval syntax (for sub-minute cadence):
Use <positive integer> seconds (e.g. 30 seconds) — the only thing 5-field cron can't express.
When to pick which: use 5-field cron for "wall-clock" cadence (every Monday at 9 AM, daily midnight, every 5 minutes on the dot). Use interval syntax when you need sub-minute cadence or simple "every N seconds" semantics. At very high cadence (e.g. 1 second), watch schedules.job_logs row counts — every fire writes a log row.
Secret References in Headers
Headers can reference secrets stored in InsForge using the syntax ${{secrets.KEY_NAME}}.
Sub-minute cadence → pg_cron interval form (e.g. 30 seconds)
6-field cron with seconds (*/2 * * * * *) is not supported — use the interval form
Store sensitive values as secrets
Use ${{secrets.KEY_NAME}} in headers for API keys and tokens
Create secrets first via the secrets API before referencing them
Target InsForge functions for serverless tasks
Use the function URL format: https://your-project.region.insforge.app/functions/{slug}
Ensure the target function exists and has status: "active"
Monitor execution logs
Check logs regularly to ensure schedules are running successfully
Look for non-200 status codes and failed executions
Common Mistakes
Mistake
Solution
Using 6-field cron (e.g. */2 * * * * *)
Not supported — use pg_cron interval form (2 seconds) for sub-minute, or 5-field cron for everything else
Referencing non-existent secret
Create the secret first via secrets API
Targeting non-existent function
Verify function exists and is active before scheduling
Schedule not running
Check isActive is true and cron expression is valid
Recommended Workflow
1. Create secrets if needed -> `npx @insforge/cli secrets add KEY VALUE`
2. Create/verify target function -> `npx @insforge/cli functions list`
3. Create schedule -> `npx @insforge/cli schedules create`
4. Verify schedule is active -> `npx @insforge/cli schedules get <id>`
5. Monitor execution logs -> `npx @insforge/cli schedules logs <id>`
Diagnose backend health
# Full health report (all checks)
npx @insforge/cli diagnose
# Check specific areas
npx @insforge/cli diagnose metrics --range 24h # CPU/memory/disk over last 24h
npx @insforge/cli diagnose advisor --severity critical # critical issues only
npx @insforge/cli diagnose db --check bloat,slow-queries # specific DB checks
npx @insforge/cli diagnose logs # aggregate errors from all sources
Authentication failures, request validation, general backend errors
Common Debugging Scenarios
Problem
Check
Function not working
function.logs
Database query failing
postgres.logs, postgREST.logs
Auth issues
insforge.logs
API returning 500 errors
insforge.logs, postgREST.logs
General health / performance
diagnose (full report) or diagnose metrics
Database bloat / slow queries
diagnose db
Security / config issues
diagnose advisor --category security
Compute service not starting
compute events <id> (machine lifecycle events)
Compute source-mode deploy failed
Verify flyctl is on PATH (flyctl version); the per-app deploy token has a 20-min TTL — re-run if expired. Use --image <url> with a pre-built image to skip flyctl entirely.
Compute image-mode deploy failed
Confirm the image is publicly pullable (private registries need per-project credential setup)
Non-interactive CI/CD
INSFORGE_EMAIL=$EMAIL INSFORGE_PASSWORD=$PASSWORD npx @insforge/cli login --email -y
npx @insforge/cli link --project-id $PROJECT_ID --org-id $ORG_ID -y
npx @insforge/cli db query "SELECT 1 AS ok" --json
Project Configuration
After create or link, .insforge/project.json is created:
oss_host is the base URL for all SDK and API operations. api_key is the full-access admin key for backend API calls, equivalent to a service role key on other platforms.
Never commit this file to version control or share it publicly.
Do not edit this file manually. Use npx @insforge/cli link to switch projects.
