MIAPI — Grounded Web Answers

uk.miapiPublished by miapi.ukv1.1.1Updated Oct 25, 2026

Web search and AI answers with real source citations and a confidence score.

VerifiedSTDIODesktop onlyWeb Search & ScrapingAI & ML

Overview

AI-generated overview

MIAPI is a local MCP server that gives an assistant web-grounded AI answers with source citations, plus web, news and image search.

What you can do
Get AI answers grounded in real-time web search, with citations and a confidence score, optionally in knowledge mode from supplied text or streamed as it is generated. Run raw web searches and return source titles and URLs. Search news by topic and image search by query. Check monthly query usage and plan tier.
When to use it
Useful when you want an assistant to answer questions from current web sources and show where the answer came from, rather than relying on model memory. Also fits quick news or image lookups and answers restricted to your own text.
Requirements
Runs as a local process installed from the miapi-sdk Python package with the mcp extra, so Python is needed. Requires a MIAPI API key supplied in the MIAPI_API_KEY environment variable; a free tier is available. Network access to the MIAPI service is required.
Before you install
The MIAPI_API_KEY secret is required and is sent to the MIAPI service. Queries, including any knowledge text you pass in knowledge mode, go to a third party. Usage is metered against a monthly limit, and paid tiers exist, so heavy use may incur cost.

Installation

In SourceWeft

  1. Open MIAPI — Grounded Web Answers in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

MIAPI — Grounded AI Answers API

Official Python client for MIAPI — get AI-powered answers grounded in real-time web search.

Install

bash
pip install miapi-sdk

Quick Start

python
from miapi import MIAPI
client = MIAPI("YOUR_API_KEY")
# Get a grounded answer with citationsresult = client.answer("What is quantum computing?", citations=True)print(result.answer)     # AI-generated answer with [1][2] markersprint(result.sources)    # List of source URLsprint(result.confidence) # 0.0 - 1.0

Features

Web-Grounded Answers

python
result = client.answer("Who won the last World Cup?")print(result.answer)

Knowledge Mode — Answer From Your Data

python
result = client.answer(    "What is the return policy?",    mode="knowledge",    knowledge="Returns accepted within 30 days with receipt...")print(result.answer)

Search Only — Raw Results

python
sources = client.search("latest AI research papers")for source in sources:    print(source.title, source.url)

News Search

python
news = client.news("technology", num_results=5)for article in news:    print(article.title, article.date)

Image Search

python
images = client.images("golden retriever")for img in images:    print(img.url, img.width, img.height)

Streaming

python
for event in client.stream("Explain quantum computing"):    if event['type'] == 'answer':        print(event['content'], end='', flush=True)    elif event['type'] == 'done':        print(f"\nDone in {event.get('query_time_ms')}ms")

Check Usage

python
info = client.usage()print(f"Used {info.queries_this_month}/{info.monthly_limit} this month")print(f"Tier: {info.tier}")

Custom Options

python
result = client.answer(    "Compare Python and Rust",    response_format="markdown",   # text, short, json, markdown    temperature=0.5,              # 0.0 (precise) to 1.0 (creative)    max_tokens=800,               # Max answer length    language="French",            # Answer in any language    search_domains=["wikipedia.org"],  # Restrict sources    system_prompt="Answer like a professor",)

Error Handling

python
from miapi import MIAPI, MIAPIError, RateLimitError, AuthenticationError
client = MIAPI("YOUR_API_KEY")
try:    result = client.answer("Hello")except AuthenticationError:    print("Bad API key")except RateLimitError as e:    print(f"Rate limited: {e.message}")except MIAPIError as e:    print(f"Error: {e.message} (HTTP {e.status_code})")

OpenAI Drop-in Replacement

MIAPI is also compatible with the OpenAI Python client:

python
from openai import OpenAI
client = OpenAI(    api_key="YOUR_MIAPI_KEY",    base_url="https://api.miapi.uk/v1")
response = client.chat.completions.create(    model="miapi-grounded",    messages=[{"role": "user", "content": "What is AI?"}])print(response.choices[0].message.content)

Get Your API Key

Sign up free at miapi.uk — 500 queries/month included.

Links

Use from Claude Desktop, Cursor or Windsurf (MCP)

MIAPI ships an MCP server, so an AI assistant can search the web and get answers with real citations without you writing any code.

bash
pip install "miapi-sdk[mcp]"

Then add it to your assistant's MCP config:

json
{  "mcpServers": {    "miapi": {      "command": "miapi-mcp",      "env": { "MIAPI_API_KEY": "your_key_here" }    }  }}

Tools it provides: web_answer, web_search, news_search, image_search, check_usage.

Get a key at https://miapi.uk — free tier included.

Source: README.md at commit 1f6137c

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v1.1.1LatestSep 16, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts