attio-cost-tuning

name	attio-cost-tuning
description	Optimize Attio API usage costs -- reduce request volume, select the right plan, monitor usage, and implement budget alerts. Trigger: "attio cost", "attio billing", "reduce attio costs", "attio pricing", "attio expensive", "attio budget", "attio usage".
allowed-tools	Read, Grep
version	1.7.0
license	MIT
author	Jeremy Longshore <[e-mail]>
tags	["saas","crm","attio"]
compatibility	Designed for Claude Code

Attio Cost Tuning

Overview

Attio pricing is based on workspace seats, not API calls. However, API rate limits effectively cap throughput, so optimizing request volume improves both performance and cost efficiency. This skill covers practical strategies to reduce unnecessary API calls.

Attio Pricing Model

Plan	Price	Key Limits
Free	$0/user/mo	3 users, basic objects, limited automations
Plus	$29/user/mo	Unlimited objects, lists, advanced reporting
Pro	$59/user/mo	Advanced automations, API access, webhooks
Enterprise	Custom	SSO, audit logs, dedicated support, custom rate limits

API access requires Plus plan or higher. Rate limits are per-workspace, not per-seat.

Instructions

Step 1: Audit Current API Usage

// Instrument all API calls to measure usage patterns
class AttioUsageTracker {
  private calls: Array<{
    method: string;
    path: string;
    timestamp: number;
    durationMs: number;
    cached: boolean;
  }> = [];

  async track<T>(
    method: string,
    path: string,
    operation: () => Promise<T>,
    cached = false
  ): Promise<T> {
    const start = Date.now();
    try {
      const result = await operation();
      this.calls.push({
        method, path, timestamp: start,
        durationMs: Date.now() - start, cached,
      });
      return result;
    } catch (err) {
      this.calls.push({
        method, path, timestamp: start,
        durationMs: Date.now() - start, cached: ,
      });
       err;
    }
  }

  (windowMs = ): {
    : ;
    : ;
    : <{ : ; :  }>;
  } {
     cutoff = .() - windowMs;
     recent = ..( c. > cutoff);
     cached = recent.( c.).;

     endpointCounts =  <, >();
     ( call  recent) {
       key = ;
      endpointCounts.(key, (endpointCounts.(key) || ) + );
    }

     topEndpoints = [...endpointCounts.()]
      .( b[] - a[])
      .(, )
      .( ({ path, count }));

     { : recent., : cached, topEndpoints };
  }
}

Step 2: Reduce Request Volume

The five biggest cost/rate-limit savers:

Strategy	Reduction	Implementation
Cache object schemas	50-90% of schema reads	Cache GET /objects and /attributes for 30 min
Batch with $in filter	N:1 on lookups	Single query instead of N individual fetches
Use limit: 500	5x fewer pagination requests	Max page size per request
Webhook-driven sync	Eliminate polling	React to changes instead of polling every N seconds
Cache records	30-80% of record reads	LRU cache with webhook invalidation

Step 3: Eliminate Polling with Webhooks

// BAD: Polling every 30 seconds for changes
setInterval(async () => {
  const records = await client.post("/objects/people/records/query", {
    filter: { updated_at: { $gt: lastCheck.toISOString() } },
    limit: 500,
  });
  for (const record of records.data) await processUpdate(record);
  lastCheck = new Date();
}, 30_000);
// Cost: 2,880 requests/day MINIMUM (even with no changes)

// GOOD: Webhook-driven (0 requests when no changes)
app.post("/webhooks/attio", async (req, res) => {
  res.status(200).json({ received: true });
  const event = req.body;
  if (event.event_type === "record.updated") {
    const record = await client.get(
      `/objects/${event.object.api_slug}/records/${event.record.id.record_id}`
    );
    await processUpdate(record);
  }
});
// Cost: 1 request per actual change

Step 4: Smart Caching Tiers

import { LRUCache } from "lru-cache";

// Tier 1: Schema data (changes very rarely)
const schemaCache = new LRUCache<string, unknown>({
  max: 100,
  ttl: 30 * 60 * 1000, // 30 minutes
});

// Tier 2: Record data (changes occasionally)
const recordCache = new LRUCache<string, unknown>({
  max: 5000,
  ttl: 5 * 60 * 1000, // 5 minutes
});

// Tier 3: List/query results (changes frequently)
const queryCache = new LRUCache<string, unknown>({
  max: 200,
  ttl: 30 * 1000, // 30 seconds
});

function getCacheForPath(path: string): LRUCache<string, unknown> {
  if (path.includes("/attributes") || path === )  schemaCache;
   (path.() && !path.())  recordCache;
   queryCache;
}

Step 5: Request Budget Monitor

class AttioRequestBudget {
  private requestsToday = 0;
  private dayStart = this.todayStart();
  private readonly dailyBudget: number;
  private readonly warningThreshold: number;

  constructor(dailyBudget = 10_000) {
    this.dailyBudget = dailyBudget;
    this.warningThreshold = dailyBudget * 0.8;
  }

  private todayStart(): number {
    const d = new Date();
    d.setHours(0, 0, 0, 0);
    return d.getTime();
  }

  recordRequest(): void {
    const today = this.todayStart();
    if (today !== this.dayStart) {
      this.dayStart = today;
      this.requestsToday = 0;
    }
    this.requestsToday++;

     (. === .(.)) {
      .();
    }

     (. >= .) {
      .();
    }
  }

  (): { : ; : ; :  } {
     {
      : .,
      : .,
      : .((. / .) * ),
    };
  }
}

Step 6: SQL Usage Dashboard

If you log API calls to a database:

-- Daily request volume (last 30 days)
SELECT
  DATE(timestamp) AS day,
  COUNT(*) AS total_requests,
  COUNT(CASE WHEN cached THEN 1 END) AS cache_hits,
  ROUND(COUNT(CASE WHEN cached THEN 1 END) * 100.0 / COUNT(*), 1) AS cache_hit_pct,
  ROUND(AVG(duration_ms), 0) AS avg_latency_ms
FROM attio_api_log
WHERE timestamp >= CURRENT_DATE - INTERVAL '30 days'
GROUP BY DATE(timestamp)
ORDER BY day DESC;

-- Top endpoints by volume (identify optimization targets)
SELECT
  method || ' ' || path AS endpoint,
  COUNT()  calls,
  ROUND((duration_ms), )  avg_ms,
  (  status     )  rate_limited
 attio_api_log
      
  , path
  calls 
LIMIT ;

Cost Optimization Checklist

[ ] Object schema calls cached (30-minute TTL)
[ ] Record lookups cached (5-minute TTL with webhook invalidation)
[ ] Polling replaced with webhooks where possible
[ ] Bulk operations use $in filter (1 request instead of N)
[ ] Pagination uses limit: 500 (max page size)
[ ] Unnecessary API calls identified and eliminated
[ ] Usage monitoring in place with daily budget alerts
[ ] Cache hit rate > 50% on read-heavy workloads

Prerequisites

Confirm that you have an Attio workspace appropriate to the task, a dedicated non-production record or workspace for testing, and only the API token scopes or administrative access required by the procedure.

Output

Following this guide produces the Attio integration outcome for its topic—configuration, validation evidence, operational recovery, or a documented migration result. Record command output and relevant identifiers so a failed step is traceable.

Examples

Start with the smallest applicable command or code example in the relevant section, using a dedicated test record or workspace and non-production credentials. Confirm the expected response or validation result before applying the pattern to production.

Error Handling

Cost issue	Root cause	Fix
High request volume	Polling loop	Switch to webhooks
Low cache hit rate	Short TTL or no cache	Increase TTL, add webhook invalidation
Rate limiting (429s)	Burst without throttling	Add PQueue with intervalCap
N+1 queries	Individual record fetches	Batch with $in filter

Resources

• Attio Pricing
• Attio Rate Limiting
• Attio Webhooks Guide

Next Steps

For architecture patterns, see attio-reference-architecture.