Semantiqv0.5.2
01Home
02Features
03Docs
04Blog
05Changelog
06Support
Get Started
  1. Home
  2. Blog
  3. Semantiq v0.6.0: HTTP API Server & Smarter Import Resolution
updates
5 min read

Semantiq v0.6.0: HTTP API Server & Smarter Import Resolution

Semantiq now offers an HTTP API alternative to MCP stdio, accurate local import resolution for JS/TS/Python/Rust/Go, Python stdlib detection, and Docker deployment support.

Semantiq Team
February 18, 2026|5 min read
Share this article
semantiqreleasehttp-apidockerimport-resolution

Semantiq v0.6.0 is here with two major capabilities: a full HTTP API server for flexible deployment, and local import resolution that improves dependency tracking accuracy.

HTTP API Server: Beyond MCP Stdio#

Until now, Semantiq only communicated via MCP (Model Context Protocol) over stdio. This works great for AI coding assistants like Claude Code and Cursor, but limits integration options.

v0.6.0 introduces an HTTP API server that exposes all Semantiq capabilities over REST endpoints:

Terminal
1# Start the HTTP server on port 8080
2semantiq serve --http-port 8080
3
4# Optional: Configure CORS for browser access
5semantiq serve --http-port 8080 --cors-origin "http://localhost:3000"

Available Endpoints#

EndpointMethodDescription
/healthGETHealth check for load balancers
/statsGETIndex statistics and calibration data
/searchPOSTSemantic code search
/find-refsPOSTFind symbol references
/depsPOSTAnalyze file dependencies
/explainPOSTGet detailed symbol explanations

Built-In Safeguards#

The HTTP server includes production-ready middleware:

  • 1MB body limit — prevents memory exhaustion from oversized requests
  • 50 concurrent requests — protects against request flooding
  • Configurable CORS — secure cross-origin access for web UIs
  • Structured JSON errors — consistent error responses for clients

Use Cases#

The HTTP API supports new integration patterns:

  1. Custom UIs — Build web dashboards for code exploration
  2. CI/CD integration — Query the index during builds without MCP
  3. Multi-language clients — Any language with HTTP can use Semantiq
  4. Microservice deployment — Run Semantiq as a standalone service

Local Import Resolution: Following the Code#

Previously, Semantiq tracked imports as raw strings. When your TypeScript file imported ./utils/helpers, Semantiq stored exactly that — the string ./utils/helpers.

v0.6.0 fixes this. Semantiq now resolves import paths to actual files on disk:

TypeScript
1// Your code
2import { formatDate } from './utils/helpers';
3
4// Old behavior: stores "./utils/helpers"
5// New behavior: stores "/absolute/path/to/src/utils/helpers.ts"

Why This Matters#

Resolved paths improve find_refs accuracy. Consider this scenario:

Plain Text
1src/
2 components/
3 Button.tsx → import { theme } from '../styles'
4 Card.tsx → import { theme } from '../styles'
5 styles/
6 index.ts → export const theme = { ... }

Previously, searching for references to theme might miss these imports because the relative paths differ. Now, both imports resolve to the same absolute path, and find_refs correctly identifies them as references to the same symbol.

Supported Languages#

Local import resolution works for:

LanguageResolution Rules
JavaScript/TypeScript.js, .ts, .tsx, index.* files
Python.py files, __init__.py packages
Rustmod.rs, lib.rs, file modules
Go.go files in the same package

Schema Migration#

The new resolved_path column required a schema update (v3 → v4). Migration is automatic and incremental — your existing index is preserved, and resolved paths are populated on the next indexing pass.

Python Stdlib Detection#

Semantiq now accurately distinguishes Python standard library imports from external dependencies:

Python
1import os # ← Standard library
2import sys # ← Standard library
3import requests # ← External dependency
4import numpy # ← External dependency

This uses a built-in list of 200+ standard library modules with binary search for fast lookups. The classification affects how Semantiq displays dependencies — standard library imports are clearly labeled, making it easier to understand a file's external dependencies.

Symbol Parent Tracking#

Symbols now track their parent context. When you search for a method, Semantiq tells you which class or struct it belongs to:

Rust
1// Before v0.6.0
2// search("process") → "process" (function)
3
4// After v0.6.0
5// search("process") → "process" (method of DataProcessor)

This context helps when exploring unfamiliar codebases. Instead of seeing isolated function names, you understand the full hierarchy.

Docker Deployment#

Semantiq is now Docker-ready with a multi-stage Dockerfile optimized for deployment:

DOCKERFILE
1# Pull the official image
2docker pull ghcr.io/semantiq/semantiq:latest
3
4# Run as a service
5docker run -p 8080:8080 -v /your/code:/workspace \
6 ghcr.io/semantiq/semantiq serve --http-port 8080

The image is Railway-ready for one-click cloud deployment. Mount your codebase, expose the HTTP port, and you have a semantic code search service.

Upgrade Guide#

Upgrading is straightforward:

Terminal
1# Update to v0.6.0
2npm update -g semantiq-mcp
3
4# Your existing index migrates automatically
5# Just run the server as usual
6semantiq serve

For HTTP API usage, add the --http-port flag:

Terminal
semantiq serve --http-port 8080

What's Next#

v0.6.0 adds infrastructure for distributed Semantiq deployments. Coming releases will build on this:

  • Team indexes — Share indexes across distributed teams
  • Cloud sync — Automatic index synchronization
  • Multi-repo search — Query across multiple codebases

The HTTP API supports integration patterns beyond MCP stdio. Let us know what you build with it.


Get started today:

Terminal
npm install -g semantiq-mcp
semantiq init
semantiq serve --http-port 8080

Questions? Feedback? Join the discussion on GitHub.

← Back to Blog

Related Posts

updatesFeatured

From v0.1 to v0.5: How Semantiq Evolved Into a Production-Ready MCP Server

The story of Semantiq's evolution from a basic 9-language MCP tool to a production-ready semantic code engine with ONNX embeddings, adaptive ML thresholds, and enterprise-grade security.

Feb 14, 202612 min read
tutorialsFeatured

I Mapped 150 Files in 5 Minutes with Semantiq MCP

How Semantiq MCP tools transform code exploration: from 2 hours of grep to 5 minutes of semantic search.

Feb 18, 20262 min read
analysisFeatured

AI Technical Debt in 2026: The Numbers Nobody Wants to Hear

AI technical debt is accelerating: 211M lines analyzed show 48% more duplication, 60% less refactoring, and 7.2% less delivery stability per 25% AI adoption.

Feb 15, 202611 min read
Semantiq

One MCP Server for every AI coding tool. Powered by Rust and Tree-sitter.

GitHub

Product

  • Features
  • Documentation
  • Changelog

Resources

  • Quick Start
  • CLI Reference
  • MCP Integration
  • Blog

Connect

  • Support
  • GitHub
// 19 languages supported
Rust
TypeScript
JavaScript
Python
Go
Java
C
C++
PHP
Ruby
C#
Kotlin
Scala
Bash
Elixir
HTML
JSON
YAML
TOML
© 2026 Semantiq.|v0.5.2|connected
MIT·Built with Rust & Tree-sitter