FINDING 01·Developers building integrations on top of SaaS platforms (HubSpot, Salesforce, PayPal)·2 sources
Poor and Inconsistent API Documentation
Developers frequently encounter documentation that is unclear, inconsistent across API versions, or simply missing for key features. This forces them to piece together information from multiple versioned endpoints or rely on outdated references, significantly increasing integration time.
Source
“Hubspot released Sequences 10 years ago, and somehow they still don't have API support. Meanwhile if I need to fetch associations between objects I have to use their V4 associations API but if I want to fetch contact lists I need to use their V1 API, and to fetch contacts themselves I sometimes have to use a V1 API and sometimes a V3 API. Who is versioning all of this”
Source
“It's 2024 and somehow none of their APIs support sorting, ordering or filtering and to do so you need to use their /search endpoint.”
FINDING 02·Developers and small teams integrating payment APIs·1 source
Inadequate Support from API Vendors During Integration
Developers report that when they hit integration blockers, vendor support is unhelpful — often suggesting technology switches rather than resolving the root issue. This leaves developers stuck for days without a working solution.
Source
“Every time we have a problem the support assistance invites us to change the implementation technology, from SOAP, to SDK, to NVP, etc.. After 3 days of working (from coding with SOAP (PHP) up to the attempt to start the SDK's examples) no solution has yet been found.”
FINDING 03·Developers using AI coding agents for API integration tasks·1 source
AI Agents Using Outdated API Docs and Deprecated SDKs
As of 2025–2026, AI coding agents (e.g. Cursor) are being used to assist with API integrations but consistently reach for stale documentation or deprecated SDK versions from their training data rather than current references. This is a recent and growing pain point as AI-assisted development becomes mainstream.
Source
“The agents kept reaching for outdated docs and deprecated SDK versions, often falling back to training data instead of reading the current API documentation. We tested this ourselves with Cursor. Across multiple iterations, 87% of runs fetched outdated reference docs using web search, and 13% of implementations used a deprecated SDK version.”
FINDING 04·Developers integrating third-party APIs (payments, e-signatures, social, SaaS tools)·2 sources
OAuth / Authentication Errors During API Integration
Developers across multiple platforms and years consistently struggle with authentication failures when integrating APIs — including bad request errors on token endpoints, incorrect signature errors, invalid client IDs, and missing PKCE parameters. These span OAuth flows, HMAC signing, and API key setups, suggesting auth configuration is a perennial and widespread pain point.
Source
“I am always getting "Incorrect Signature" error when I test the application on device. However this twitter developer application keys (API and SECRET) are working fine in following cases 1) If i test the application on simultar 2) If i use these keys with other application.”
Source
“Authorization failed because the code challenge parameter is missing. PKCE is required for this authorization request. Please contact the app developer.”
FINDING 05·Developers building integrations on SaaS platforms (HubSpot, Salesforce, Aqara)·2 sources
Missing or Restricted API Features on Third-Party Platforms
Developers discover that key capabilities they need are either not exposed via API at all, or are locked behind org-level settings they cannot control. This blocks integration projects entirely and forces workarounds.
Source
“We cannot connect to them through HTTP API. Also, for connection to the Alexa thermostat we also have the problems because we cannot enable the thermostat. We do not the output thermostat device like Air Conditioner to connect, so we need to find some imitation. So far we are not successful.”
Source
“"The REST API is not enabled for this Organization," Here's the API request in question: {instance_url}/services/data/v52.0/query/?q=SELECT%20Id%2C%20Subject%2C%20Status%2C%20Priority%20FROM%20Task”
FINDING 06·Backend/cloud developers building serverless architectures on AWS·3 sources
Cloud Service Integration Misconfiguration (API Gateway, Lambda, DynamoDB)
Developers integrating AWS services — API Gateway with Lambda, DynamoDB, or Firehose — frequently encounter silent failures, missing triggers, and cross-account permission issues that are difficult to debug. Infrastructure-as-code tools like Terraform can produce configurations that appear correct but behave differently from manual setups.
Source
“If I manually add an Integration Request of type Lambda function, an Api Gateway trigger is automatically added to the lambda function. If I do it via Terraform, everything looks correct but when I go look at the Lambda function it has no trigger.”
Source
“If both are in the same accounts, this works just fine, I've had no problems with the integration. If they're in separate accounts, I can't figure out how to signal to API Gateway that the dynamo table is in a different account.”
Source
“The Terraform API Gateway integration with Firehose below works if Firehose is created separately in advance.”