Skip to main content
Product Config

Tune algorithms live

Adjust ranking weights, thresholds, and parameters in real-time

Store ranking weights and algorithm parameters in Replane. Data scientists iterate on algorithms without waiting for deploys.

The Problem & Solution

Before & after Replane

See how teams transform their workflow

Without Replane

  • Hardcoded parameters

    Algorithm weights buried in code. Changing them requires a deploy.

  • Slow experimentation

    Testing new parameter values takes days with deploy cycles.

  • Engineering bottleneck

    Data scientists wait for engineers to update and deploy parameter changes.

  • Can't respond to feedback

    Users complain about search results. Fixes take too long.

With Replane

  • Externalized parameters

    Store all algorithm parameters in Replane. Change without code.

  • Instant iteration

    Adjust weights, see results immediately. Iterate in minutes.

  • Self-service for data teams

    Data scientists tune algorithms directly. No engineering bottleneck.

  • Quick response

    Bad search results? Adjust parameters and fix in seconds.

Key Features

Iterate at the speed of thought

Empower data teams to tune algorithms directly

Instant Apply

Changes take effect in under a second

No Deploys

Update product config without code changes

Unlimited Iterations

Tune parameters as often as needed

Instant Rollback

Revert any change with one click

How It Works

Three simple steps

Get up and running in minutes, not hours

1

Define parameters

Store ranking weights, thresholds, and tuning parameters in Replane. Use structured JSON for complex configs.

2

Observe behavior

Monitor search quality, recommendation CTR, or matching accuracy. Identify opportunities.

3

Tune and iterate

Adjust parameters and see immediate impact. Find the optimal values through rapid iteration.

Benefits

Why teams choose Replane

Built for developers who value speed, reliability, and control

Rapid iteration

Test new parameter values in minutes, not days. Accelerate algorithm improvement.

Safe experimentation

Every change is versioned. Made things worse? Rollback to previous values instantly.

Measurable impact

A/B test different parameter sets. Measure the impact on your key metrics.

Code Examples

Works with your stack

Official SDKs for all major languages and frameworks

1import { Replane } from '@replanejs/sdk'
2
3interface Configs {
4 'search-weights': {
5 title: number
6 description: number
7 tags: number
8 recency: number
9 }
10 'similarity-threshold': number
11 'max-results': number
12 'boost-premium': number
13}
14
15const replane = new Replane<Configs>()
16
17await replane.connect({
18 baseUrl: 'https://replane.example.com',
19 sdkKey: process.env.REPLANE_SDK_KEY
20})
21
22function rankSearchResults(query: string, items: Item[]) {
23 const weights = replane.get('search-weights')
24 const threshold = replane.get('similarity-threshold')
25 const maxResults = replane.get('max-results')
26 const premiumBoost = replane.get('boost-premium')
27
28 return items
29 .map(item => ({
30 item,
31 score:
32 weights.title * scoreTitle(query, item.title) +
33 weights.description * scoreDescription(query, item.description) +
34 weights.tags * scoreTags(query, item.tags) +
35 weights.recency * scoreRecency(item.createdAt) +
36 (item.isPremium ? premiumBoost : 0)
37 }))
38 .filter(r => r.score >= threshold)
39 .sort((a, b) => b.score - a.score)
40 .slice(0, maxResults)
41}
FAQ

Questions about Product Config

Common questions and answers

Ranking weights, similarity thresholds, relevance boosts, matching criteria, score normalizations, and any numeric value that affects algorithm behavior. If you tune it during development, it belongs in Replane.

Yes! Give your data team dashboard access. They can adjust parameters, run experiments, and iterate on algorithms without engineering involvement. RBAC ensures they can only modify what they should.

Create multiple configs with different parameter values. Use override rules to serve different configs to different user percentages. Compare metrics in your analytics tools.

Every change is versioned with full history. If a parameter change causes issues, rollback to the previous value with one click. Changes revert in milliseconds.

Replane pushes updates to all connected clients via SSE simultaneously. All services receive new parameters at the same time, ensuring consistent algorithm behavior.

Get Started with Product Config

Ready to ship faster?

Start using Replane in minutes. Free forever for self-hosted, or let us manage everything for you.

No credit card required. MIT licensed. Deploy anywhere.