Skip to content
Sun Software logo with the tagline Engineering Beyond Code
AI API & MCP Integrations

AI integration with the tools your business already uses.

Our AI integration services connect AI with your APIs, workflow platforms and tool interfaces, so assistants can retrieve context or take approved actions inside real systems.

service-briefScoped proposal
  • CategoryAI Agents & Automation
  • EngagementScoped proposal
  • First stepDiscovery call
  • Reply withinOne business day
01

50+

Countries our clients come from

Mostly the US, UK, Germany, Canada and Australia.

02

1,685

Reviews on Fiverr

Public, and open for anyone to read.

03

16

People on the team

Working together since 2020.

04

4.9

Out of 5 on Fiverr

Average across all reviews.

AI can reach your systems. Who else can?

Connecting AI to your software means keys, permissions and connectors. Security research shows where they leak.

  • keys for AI services found exposed on public GitHub in 2025

    1.27M

    What we do

    We keep every key for your AI and your data in a vault, never in the code.

  • secrets found sitting in MCP configuration files

    24,008

    What we do

    We use official connectors, fix their version and keep the keys they need in a vault, not in a config file.

  • of keys leaked in 2022 still worked four years later

    64%

    What we do

    We change your keys on a schedule and switch the old ones off.

  • organisations had an API security incident in the past year

    1 in 3

    What we do

    We give each connection only the access its task needs, read-only where possible, with approval before changes.

  • of developers name runaway API calls from AI agents as a top risk

    51%

    What we do

    We log every call, set limits on usage and cost, and alert you when something looks unusual.

What a safe AI integration needs.

Five problems businesses meet when they connect AI to their systems, and what we do about each.

  1. Problem 01 · Keys

    Keys that unlock your AI and your data end up in code, config files and chat.

    What we do

    We keep every key in a proper vault, out of the code, and change them on a schedule.

    • Keys out of the code
    • Stored in a vault
    • Changed on a schedule
    • Old keys switched off
  2. Problem 02 · Permissions

    The assistant is given far more access than the job needs.

    What we do

    Each connection gets only the access its task needs, and actions that change data wait for approval.

    • Least access per task
    • Read-only where possible
    • Approval before changes
  3. Problem 03 · Hidden instructions

    An email, ticket or web page the assistant reads can carry instructions that misuse its tools.

    What we do

    We treat everything the assistant reads as untrusted, and test the connection with attack examples.

    • Outside text treated as data
    • Tested with attack examples
    • Unusual actions blocked
  4. Problem 04 · Connectors

    Connectors from public registries are trusted without review, and a bad update can copy your data.

    What we do

    We use official connectors, fix their version, and read the code before it touches your systems.

    • Official sources only
    • Versions fixed
    • Code reviewed first
  5. Problem 05 · Monitoring

    Nobody watches what the AI calls or how often, so cost and incidents go unnoticed.

    What we do

    We log every call, set limits on usage and cost, and alert you when something looks unusual.

    • Every call logged
    • Limits on usage and cost
    • Alerts on unusual activity

Real connectors. Real leaks.

Three public cases show how AI connections go wrong. None of them is our client.

  • Malicious package · September 2025

    A fake connector that copied every email

    What happened

    A package posing as Postmark's MCP server built trust over 15 versions. Then one update quietly sent a copy of every outgoing email to an attacker. Postmark itself was not breached.

    What we do about it

    We use official connectors, fix their version and review updates before they run.

    Source: The Register
  • Vendor bug · June 2025

    One company's data visible to another

    What happened

    A bug in Asana's experimental MCP server could have shown some of one customer's data to users in other organisations. Asana took it offline and fixed it. No misuse was reported.

    What we do about it

    We test that each connection can reach only its own data, and keep logs so exposure can be checked.

    Source: The Register
  • Security research · May 2025

    A planted issue that pulled private code

    What happened

    Researchers showed that a malicious issue in a public repository could make an AI agent using GitHub's MCP server publish private repository data. It was a demonstration, not a reported theft.

    What we do about it

    We limit each key to what the task needs and keep approval on for actions.

    Source: Invariant Labs

Our own work

Live products we built. Open them and judge the work for yourself. Work for private clients is shared on a call, with the client's agreement.

See all case studies

Connected. And under control.

What every AI connection we build comes with.

  • Keys kept safe

    In a vault, out of the code.

  • Least access

    Each connection sees only what it needs.

  • Approval for actions

    Changes to your data wait for a yes.

  • A log of every call

    What was asked, and what was done.

  • Cost limits

    Usage capped before it becomes a bill.

  • Free to change model

    Built so you can switch AI provider later.

Why choose Sun Software as your AI integration company?

The connection is the easy part. Keeping it safe is the work.

  • What usually happensKeys are pasted into a config file and forgotten.

    Keys never sit in code

    We store them in a vault and change them on a schedule.

  • What usually happensA connector is installed from a public list because it was first in the search.

    We read connectors before using them

    Official sources, fixed versions, code reviewed.

  • What usually happensThe assistant gets full access so the demo works.

    Actions wait for approval

    The AI can read what it needs. Changing your data needs a person.

  • What usually happensThe first sign of a problem is the invoice.

    You see every call and its cost

    A log and a usage limit come as standard.

In their words, not ours.

No public review names this service yet. These are recent reviews about how we work, word for word from Fiverr.

  • Great....very professional, delivered on time and exactly as I needed. Exceptional. I always find Sunsoft very easy to work with, always helpful and I'll continue using their services for the foreseeable future. No need to go anywhere else.
    Client in United KingdomOctober 2026 · Fiverr review
  • Excellent experience from start to finish. Communication was great, responses were always prompt, and the work was completed quickly. They captured the exact look and feel I wanted on the first try and delivered excellent value for the money. I would gladly work with them again and highly recommend their services.
    Client in GermanySeptember 2026 · Fiverr review
  • Choosing SunSoft was exactly the right decision. Their professionalism, extensive experience, and attention to detail have really impressed me so far. It was a complex project that exceeded my expectations! I look forward to continuing to work with SunSoft!
    Client in CzechiaAugust 2026 · Fiverr review

Questions about AI integrations

What do AI integration services include?

The connection between AI and your systems, and the controls on it. Keys are kept in a vault, each connection gets the least access it needs, changes to your data wait for approval, every call is logged, usage is capped and you stay free to change model.

Is MCP required for every AI integration?

No. We use it where it simplifies access for compatible tools. A direct API is sometimes the better choice.

How do you handle credentials?

With least-privilege access and a secrets manager. Credentials never go into AI prompts.

What decides the price of AI integrations?

Mainly systems, actions and controls. We agree the scope and the price in writing before work starts.

What decides the cost of an AI integration?

An integration is priced by what it connects and how carefully it must be controlled. Three things decide it. Systems: How many APIs or tools are connected, and how well they are documented. Actions: Whether the AI only reads data or also changes it. Controls: The permissions, logging and limits the risk calls for.

What happens after I contact you?

We reply within one business day. Then the work runs in 5 steps: Define, Secure access, Build, Test, Monitor. Each step ends with something you can review.

  1. 1Review an Integration
  2. 2We reply within one business day
  3. 3Discovery call
  4. 4Step one: Define

Ready to talk about AI integrations?

Tell us which systems need to talk to each other and what an approved outcome looks like.

  • The first conversation is free
  • NDA before you share anything
  • A reply within one business day