Best Software Engineering & DevAgents AI Agents
Autonomous coders, terminal copilots, automated test engineers, and architectural copilots. Explore our curated selection of top-performing autonomous agents and tooling in this sector, compare pricing and features, and deploy the ideal agent for your stack.
KushoAI
KushoAI adds testing and monitoring agents to a team's CI/CD process. They maintain checks and flag failures before a change reaches production.
Effie
Effie combines drafting, outlining and mind mapping in an offline-capable writing app. Documents sync between devices and its assistant can condense or expand a passage.
LoreKeeper
Between sessions, LoreKeeper remembers the tabletop campaign. During play, its AI game master narrates scenes, rolls dice and responds to what the characters do.
MetaGPT
MetaGPT assigns software-team roles to language models. A short requirement becomes plans, documents and code as the agents hand work to one another.
Epsilla
Epsilla helps vendors build and host specialist agents with retrieval over customer knowledge. It offers multi-tenant management and hosted, private-cloud or on-premises deployment.
SlideSpeak
Text and shapes stay editable when SlideSpeak converts a PDF, document or spreadsheet into PowerPoint. Its Charles assistant can revise individual slides or summarize the source.
GaiaHub AI
GaiaHub provides shared components for small LLM apps. Users can connect existing blocks or create their own and run the resulting app.
Turbotic
Turbotic turns a written procedure into automation that calls business APIs. Operations teams can inspect the workflow in production and coordinate agents with software robots.
keevx
Keevx turns a product link, photo or script into a presenter video with lip sync. Its other tools include video translation and avatar cloning.
AgentsLed
AgentsLed points to AgentLed's meeting-booking service. Its agents research prospects, prepare outreach for approval and use later call notes to inform follow-up.
assistant-ui
assistant-ui supplies open-source React and TypeScript components for building a streaming chat interface with tool calls and persistent conversations.
Viral AI Video Maker
Viral AI Video Maker offered editable templates for vertical TikTok, Reels and Shorts videos. The service appears inactive and the original domain has been unavailable.
Mindy AI
Mindy's Studio let teams generate or resize product images directly in Slack. A sketch could become a mockup and a product shot could become a short clip. The team has since joined WorkWhile.
Kubiya
Kubiya runs agents for engineering operations, including debugging build pipelines, managing deployments and working with infrastructure as code.
TestSprite
After a release, TestSprite builds and runs end-to-end checks from a CLI, browser, CI system or MCP-connected IDE and reports the regressions it finds.
Latta AI
Session replay gives Latta AI the steps a user took just before a defect appeared, evidence its debugging agent uses to propose a patch for review.
SageFlow
Sageflow lets teams build agents on a drag-and-drop canvas, connect tools and publish or discover ready-made agents in a marketplace.
Bloop AI
Bloop AI offered codebase questions, regex and symbol search over indexed repositories. The company ceased operating in 2026. Its source remains available but unmaintained.
Browser Use
Browser Use provides an open-source browser agent and hosted browsers for tasks that need a real session, such as logging in and extracting structured results.
Owlity
Owlity records a test and runs it in the cloud. Product and engineering teams can create and manage automated checks without writing test code.
Synthesia
Training teams can give Synthesia a script, document or slide deck and get a narrated video with an AI presenter and motion graphics.
Momentic AI
Mo, Momentic's testing assistant, captures the steps that expose a bug in a web, iOS or Android app and turns the run into a regression test for future releases.
Rig
Rust developers can call several model providers and define agent tools through Rig's type-safe interface.
GptPanda
GPT Panda adds a writing and research assistant to Slack channels and direct messages. Ask it to summarize a thread or brainstorm a reply.
CodeGPT
CodeGPT is a coding agent for VS Code, JetBrains and Visual Studio. Developers connect their own model keys for refactoring, debugging and generation.
Cotester
CoTester creates and runs software tests for enterprise applications. Its test cases are designed to adapt as the application changes.
Airtop AI
A written procedure in Airtop can run on a schedule across APIs and signed-in websites. Examples include reconciling invoices and enriching lead records.
StackAI
StackAI connects internal data and tools on a visual canvas, where teams build, deploy and govern agent workflows that retrieve company knowledge.
Cursor
Inline suggestions in Cursor cover the line being edited, while its coding agent can follow a change across the repository. Developers review the proposed edits in the same editor.
BaseAI
BaseAI now points to Command Code, a coding agent for developers using open models. It supports tool calls from a desktop interface or the command line.
AskYourDatabase
AskYourDatabase answers database questions with a chart and the SQL that produced it. Use the desktop app with your own database or embed a chatbot on a site.
Inltayer
The Inltayer listing points to Intlayer, a JavaScript internationalization toolkit. Developers can manage translations per component in React, Next.js, Vue and Svelte.
AI Agent Token Cost Calculator
Enter token counts, model prices and run frequency in the AI Agent Token Cost Calculator to estimate a monthly bill. Large logs and noisy commands can be included in the comparison.
EducationAds AI
EducationAds AI helps enrollment teams review Meta ads and develop new hooks, scripts and campaign concepts using examples from the education market.
Genesis Bots
A schema, its lineage and the rules attached to it guide Genesis Bots on Snowflake and Databricks. Data teams use the agents to investigate failed pipelines or help migrate legacy systems.
GPT Researcher
GPT Researcher is an open-source deep research agent. It breaks a question into research tasks, searches the web or local files and assembles a report with citations. The Python package uses your own model and search keys.
QodoAI
A change can satisfy a local check and still conflict with an earlier requirement. QodoAI reads the surrounding code and past pull requests when reviewing a proposed edit in a PR or IDE.
ShortGPT
From a script to a narrated short, ShortGPT handles footage selection, voiceover and editing. The open-source framework includes an editing language designed for language models.
Taletok
Taletok creates narrated or illustrated videos for faceless YouTube channels from a topic, with formats for full-length episodes and vertical Shorts.
Agent Oracle
Signed receipts make Agent Oracle's decisions independently checkable. An agent submits a factual claim before acting. A third party can later verify the decision offline.
Confident AI
Confident AI lets teams compare evaluations with live application traces and expert feedback to track where model or agent quality changes.
WorkerGen
WorkerGen turns a short business brief into social captions, layouts and visuals. It checks the assets against brand guidance before export.
Fine
A GitHub, Linear or Jira issue can go into Fine and return as a pull request. Its coding agents plan and implement the change in a cloud environment.
Magicley AI
Magicley AI combines chat, image generation, voice and custom bots in a credit-based workspace. Its conversational interface is called Abra-Chat-Dabra.
Tabby
For teams that host their own development tools, Tabby supplies open-source code completion and answers to coding questions within their environment.
TRAE
TRAE offers both a code editor and a separate workspace for longer agent tasks. It can import VS Code or Cursor settings on macOS, Windows and Linux.
Buildel
Buildel lets users connect visual blocks for chat, document search, audio and browser actions. The no-code agent builder works with the user's own model keys.
Sweep AI
In a JetBrains project, Sweep AI can trace a problem through the code, propose edits, run the tests and draft a commit message for the resulting diff.
Komment
Komment monitors repositories for application-security findings and technical debt. It covers projects where coding agents contribute changes, producing risk reports and suggested fixes.
Continue
Continue was an open-source project for configuring coding agents and coordinating them through a graphical interface. Cursor acquired the project. Its open-source code remains available.
Replit Agent
A prompt in Replit Agent can become a published app in the same cloud workspace, with authentication, a database and hosting available as the project takes shape.
NOFire AI
On-call teams can inspect the actions NOFire has already taken on an incident. Its agents follow team policies when responding to production issues and support tickets.
gpt-pilot
gpt-pilot is an open-source coding agent that turns a written app idea into interface components and application logic, then works through errors in the codebase.
BAML
BAML defines typed functions for model output and carries the same types into production. Boundary's toolchain includes testing, tracing and alerts.
CreatOK
CreatOK helps a TikTok Shop seller study a shoppable video's hook and call to action, then produce a clip featuring their own product.
Nexus Agent
The starting material for Nexus Agent is a procedure staff already follow. Teams describe the steps, and an agent carries them out in sequence.
Hooked
Hooked teaches a method for scripting UGC ads, from the audience and angle to the opening hook. Its optional studio uses the buyer's own model keys.
CallFluent AI
Businesses can put a CallFluent AI agent on a phone number or offer a call from a web page. Its no-code setup covers incoming support questions as well as outbound reminders and surveys.
Maige
Maige uses plain-English rules to label, assign and review GitHub issues and pull requests. Its source is available for developers to inspect. New signups are closed.
Dosu
Dosu records the reasoning behind a change as part of the codebase's documentation, giving developers and coding agents shared project memory for later sessions.
Stockimg.ai
Stockimg.ai generates and edits marketing visuals such as logos, posters and stock photos. Finished posts and short videos can be scheduled to social accounts.
Sourcegraph Cody AI
The Sourcegraph entry covers codebase search and context for developers and coding agents. Its code graph connects related code, and batch changes can span repositories.
Decipher AI
Application code, user paths and API contracts give Decipher AI the material for its tests. The testing agent flags flaky checks and repairs selectors when the interface changes.
Supermaven
VS Code, JetBrains and Neovim users can get Supermaven completions in the editor they already use. Its suggestions draw on project context, with a chat panel for follow-up questions.
TwinTone AI
On a TwinTone AI livestream, a virtual host presents items from a Shopify catalog and answers viewer questions, with highlights available to clip afterward.
Imbue
Imbue develops personal software projects, including an extensible coding workspace, tools for self-hosting software and a filter for an X feed.
Windsurf
Windsurf is listed here under its earlier name. Cognition's Devin Desktop lets developers move a coding-agent session between a local machine and Devin Cloud with its context intact.
EntelligenceAI
Entelligence AI reviews pull requests and tracks how coding agents reach models and tools, including what each route costs.
Hugo
Unresolved tickets leave Hugo with their conversation transcript attached. For routine requests, the support agent draws on the team's docs and helpdesk, with MCP connectivity and a choice of model.
Discover Gift Ideas
Answer seven questions about a recipient and Discover Gift Ideas suggests eight gifts with Amazon links. Shared wishlists help people coordinate purchases.
Amplify Security
Amplify Security lets engineers run detection agents on their own repositories and build pipelines. Findings can move from triage to fixes in the same console.
Tracio
Tracio assesses a visitor's device, network and behavior as a page loads. Its verdict lets a site allow, challenge or deny the request.
CodeRabbit
Large pull request on a busy review day? CodeRabbit summarizes the diff, comments on specific changes and ranks the queue by risk.
Baz
Security, concurrency, performance and style each get a separate review in Baz. Its agents assess a pull request against surrounding code and the team's previous feedback.
Chat4Data
Tell Chat4Data which fields you need from a website. The Chrome extension follows the pages and exports the rows, pausing when you need to sign in.
Alloy
Alloy lets a product manager capture an existing web app, describe a change and share a sandbox preview for comments. Approved changes can become GitHub pull requests.
Tempo
In Tempo, a customer issue or product requirement can become a pull request. Agents implement the change in a shared IDE where the product and engineering teams review it.
Capalyze
Capture a page in Chrome or Edge, then ask questions about its rows. Capalyze turns the page into a spreadsheet and uses the captured data for charts and a short report.
Posium
Posium maps a web or mobile app before generating automated tests for it. Separate agents handle discovery, planning and test code.
Biel.ai
Readers can search technical docs in a website chat, while coding agents reach the same material through MCP. Biel.ai links answers back to the source page.
AutoQA
AutoQA takes test plans in plain English, runs them against an interface and files structured bugs in Jira when a check fails.
Breadcrumb.ai
Sponsorship renewals depend on what the partner actually received. Breadcrumb.ai compares promised inventory with campaign delivery across sports, events, media and hospitality.
Vibeknow
Course notes, PDFs and Notion pages can all become narrated explainers in Vibeknow, with the script and cut open for review before publication.
Kadoa
Describe a dataset to Kadoa and its agents build a public-web data pipeline for it. Investment teams can trace individual values to the source.
AgentLed
AgentLed researches accounts and drafts outreach inside tools such as Clay, Apollo and HubSpot. A person approves the message before delivery.
KaneAI
KaneAI lets testers describe a web or mobile check in natural language and run it from their project tools or terminal.
Edge Arena
Edge Arena asks competing agents to tackle the same business question. It scores the strategies and shows the reasoning behind the resulting plan.
SoftGen
SoftGen turns a brief into a full-stack web app with a managed Supabase backend. Founders can export the repository to GitHub or deploy to Vercel.
Dawn
Ask Dawn how a system works from Slack, Teams or Telegram. Its answers draw on the connected codebase and infrastructure.
Cobalt
Every coding task gets its own cloud computer in Cobalt, with a choice of agent and machine profile for one-off assignments or recurring jobs.
PureCode AI
PureCode AI connects a VS Code session to coding agents for .NET, Java and C++. A change can start in chat and end as a tested pull request.
Bismuth
Bugs surfaced by Bismuth can arrive as pull requests with proposed fixes that take the team's existing repository and code patterns into account.
Kane CLI By TestMu AI
Kane CLI takes a test step written in English, runs it with Playwright and returns a pass or fail with evidence. Tests can cover a browser, an API or a device.
Snipara
A decision made in Claude Code can still be found when a developer moves to Cursor or Codex. Snipara keeps cited project memory for coding agents and checkpoints for interrupted runs.
OpenHands
OpenHands runs coding agents against a repository in a sandbox. Choose a model, then run tasks locally or through a managed deployment.
Blackbox AI
Blackbox offers infrastructure for training and running models on GPUs. Its agents monitor machines and tune training against benchmarks you choose.