Developer Program — Earn while you build

Build on Signallloom.
Earn on Everything.

The Signallloom Developer Toolkit turns every AI feature you build into a revenue stream — not a cost center. Route AI traffic, earn passively, refer clients. Three ways to earn. All of them real.

Start Building — Free See the Tools
signalloom-developer
$ npm install @signalloom/costestimate
costestimate — AI cost estimation before you build
$ npm install @signalloom/sdk
SDK — route AI calls and earn 25% passive revenue
$ costestimate --model gpt-4o --prompt "summarize this"
{ "model": "gpt-4o", "costPerCall": 0.002, "monthlyEst": "$120" }
▲ gpt-4o-mini: $0.005/call — 75% cheaper for this task
$ signalloom refer --tool jobaudit
Refer clients to jobaudit. Earn $18–$1,875 per referral.

Four tools.
Each one earns. Together, they compound.

Every tool in the Signallloom toolkit has a clear ROI on its own. Used together, they create outcomes that no single tool can produce alone.

$
costestimate Save before build

Tells you exactly what any AI feature will cost per call, per user, per month — before you write a line of code. Compare models, find the cheapest option that meets your requirements, and scope features with real numbers instead of guesses.

# Estimate any AI call cost before building
costestimate --model gpt-4o \
  --prompt "classify this ticket"
# → $0.002/call · gpt-4o-mini: $0.0003 (85% cheaper)
jobaudit Find hidden waste

Runs a free audit on your AI cron jobs and production pipelines. Finds overspec problems — jobs still running GPT-4o when GPT-4o-mini would do the same job for 95% less. Shows you exactly what you're paying too much for.

signalloom jobaudit
# Analyzing 12 jobs...
# ⚠ job-7: gpt-4o @ $0.08/call
  → swap to gpt-4o-mini: $0.005/call
  Savings: $74.20/month
🔄
promptcache Reduce token waste

Analyzes your prompt structure and identifies cache opportunities. Shows which jobs share base prompts, what prefix tokens you can cache, and how much of your token bill is pure redundancy. Finds the waste that model swaps miss.

promptcache analyze
# 4 jobs share 780-token base prompt
# Est. cache hit rate: 72%
Prefix caching: save 780 tokens/call
Additional savings: $0.004/call × 2K calls = $8/day
📋
contextbroker Institutional memory

Persistent storage for your team's AI cost decisions. Every costestimate run, every model choice, every tradeoff — stored and retrievable by any agent or team member. Your AI cost knowledge stops living in Slack threads and starts compounding.

contextbroker set team/model-decisions/classifier
{ "model": "gpt-4o-mini", "reason": "simple classification",
  "date": "2026-04-06", "alternativesConsidered": [...] }
Decision stored. retrievable by any agent.

The real leverage is in the layers.

Every tool is useful alone. But when they feed each other, reinforce each other, and store what they learn — that's when the outcomes no single tool can produce start appearing.

costestimate + contextbroker

Plan before you build. Remember what you decided.

costestimate finds the right model at build time. contextbroker stores the decision permanently. Future you — or a new developer — never has to re-litigate old choices. Every decision becomes institutional knowledge.

$0.08/call GPT-4o $0.005/call GPT-4o-mini — 94% savings, remembered forever
jobaudit + promptcache

Find the overspec. Find the waste. Fix everything at once.

jobaudit tells you what you're overpaying for. promptcache tells you whether the fix is a model swap or a caching opportunity — or both. Alone, each finds partial savings. Together, they find the full picture.

$80/month $3/month — 96% total reduction, not found by either tool alone
promptcache + contextbroker

Cache analysis that persists across every session.

promptcache identifies what's cacheable. contextbroker stores the cache strategy. When a new agent or new session runs the same job, it reads the config automatically — no re-analysis needed. Your caching isn't a point-in-time report. It's living infrastructure.

One-time analysis → Persistent 72% hit rate across all agents, all sessions
costestimate + jobaudit

Stop overbuilding. Audit what's already burning money.

costestimate prevents overspec at build time. jobaudit finds it in production. You save twice — once before the feature ships, once in your existing jobs. The compound effect: you're not just auditing, you're building with cost awareness baked in.

$800/month AI bill $15/month — right-sized at build time + fixed in production

Three ways to earn.
All of them real.

This isn't a referral link you paste in a README. These are the actual economics of the platform, based on the actual pricing.

1

Build with cost awareness. Earn on every AI call.

Use costestimate before you build. Find the cheapest model that meets your requirements. You catch overspec before it ships — your features make margin instead of burning it. For every AI feature you ship with the Signallloom SDK, you earn 25% of every dollar spent routing through it.

Your cut: 25% of all AI spend routed through your SDK integration
# 200 customers × $2/month AI routed = $400/mo spend
Your passive cut: $100/month
# For as long as the feature runs. Forever.
2

Refer clients to jobaudit. Earn per referral.

Point your clients at signalloomai.com/jobaudit. We run a free audit, find their overspec problems, and show them exactly what they'd save. If they approve the changes, Signallloom charges 25% of the first three months' savings — and you get 25% of Signallloom's fee.

Earn: $18–$1,875 per referral
# Average client saves $500/month
# Signallloom earns: $375 (25% × $500 × 3mo)
Your cut: $93.75 per referral
# 2 referrals/month = $187.50 passive on top
3

Use costestimate for every build decision.

Every feature you scope with costestimate is a feature built to pencil out. You're not guessing at margins — you're making data-backed routing decisions. That means more features ship profitable, fewer features go negative, and your AI bill stays predictable as you scale.

Save before you build: $0.08/call → $0.005/call
# GPT-4o for simple classification: $0.08/call
# GPT-4o-mini for same task: $0.005/call
# At 10K calls/month: $800 → $50 = $715/month saved

Revenue share — real math

These aren't projections. These are the actual economics of the platform based on actual pricing.

Scenario AI Volume Your Cut Monthly Earnings
100 customers × $2/mo AI routed $200/month 25% $15/month passive
500 customers × $5/mo AI routed $2,500/month 25% $625/month passive
1 referral — avg $500/mo savings client Signallloom earns $375 25% $93.75 per referral
2 referrals/month at average $187.15/month referral income
Combined: 500 routed + 2 referrals/mo $812.15/month total

Run the numbers against your user base. Then run them against your client list.

Start free.
Scale when it makes sense.

No surprise bills. No character limits. Predictable pricing that grows with your usage.

Free
$0 /mo
For developers evaluating the toolkit
  • costestimate (limited calls)
  • SDK access
  • jobaudit (1 audit/month)
  • Community support
Loom Partner
$19 /mo
For indie developers and small teams
  • Unlimited costestimate calls
  • SDK with 25% revenue share
  • 5 jobaudit scans/month
  • promptcache analysis
  • contextbroker (1MB storage)
  • Email support
Enterprise
Custom
For teams with volume requirements
  • Everything in Loom Elite
  • Custom SDK integration
  • Dedicated account manager
  • Custom revenue share tiers
  • SSO / SAML
  • Invoice billing / Net-30

Bot-to-bot pricing? Ask about Loom Partners — up to 28% revenue share for AI platforms.

Developers who earned while they built

Real outcomes from real Signallloom developers. (Case studies in development — submit yours to be featured.)

I scoped an AI feature using costestimate before writing any code. It showed me GPT-4o was 16× more expensive than needed for the task. I built it right the first time. The feature went live profitable instead of going negative after launch.

Developer, SaaS Startup Built with costestimate — saved before shipping

I referred two clients to jobaudit in the same month. One was burning $340/month on overspec'd cron jobs. After the fix, their bill dropped 91%. I earned $187.50 for two emails.

Freelance Developer Referral earnings — jobaudit program

The 2+2>4 effect is real. Using jobaudit + promptcache together found $340/month in production savings. Using costestimate + contextbroker built a team knowledge base that means we never re-litigate an old model decision. We save money and move faster.

Engineering Team, Series A Full-stack toolkit users — compound savings

Have a Signallloom Developer story?

Submit your case study →

Free to try.
Real earnings from day one.

Install costestimate. Run it on your current project. In five minutes you'll know whether your AI features are built to make margin — or built to burn it.

signallloom-quickstart
$ npm install @signalloom/costestimate
@signalloom/costestimate installed
$ costestimate --model gpt-4o --prompt "summarize this"
{ "costPerCall": 0.002, "monthlyEst": "$120" }
▲ gpt-4o-mini: $0.0003 — 85% cheaper for this task
$ npm install @signalloom/sdk
@signalloom/sdk installed — earn 25% on all routed calls
$ signalloom jobaudit
Refer clients. Earn $18–$1,875 per referral.
Start Building — Free

No credit card required  ·  Free tier includes costestimate + 1 jobaudit/month