opens in a new tab

Last updated on October 26, 2026

AI tools for developers

We organize AI tools by the job they do and the limits in their docs, with our five-tool coding and model stack and six documented cross-tool workflows.

October 2026 releases

Hot this month

Model and tool updates that open up a new workflow or local capability, with the caveats to check before you switch.

  • Gemini 4 Argon rolls out in Google Antigravity

    What changed
    Google rolled out Gemini 4 Argon in the Antigravity desktop app and agy CLI (with current model docs also listing IDE extensions) for Google AI Pro, Ultra, and eligible Gemini Enterprise users.
    Why it matters
    Developers using Antigravity for architecture, debugging, or multi-file refactors can switch to Argon inside their existing terminal, desktop, or editor session.
    Limits to check
    Unavailable in the standalone Antigravity IDE. Requires paid Google AI Pro, Ultra, or eligible Enterprise access (or an eligible device Pro trial) and plan limits.
    Antigravity Gemini 4 Argon releaseopens in a new tabDetails

  • Comfy Agent launches in Comfy Cloud for node-graph workflows

    What changed
    Comfy Agent plans, builds, runs, and edits visible ComfyUI node graphs directly on the canvas alongside you, with support for reusable public or private skills and parallel chats.
    Why it matters
    Instead of handing back a single opaque asset, the agent leaves an inspectable node pipeline—models, seeds, samplers, and masks—that teams can tweak and rerun for product variants.
    Limits to check
    Beta on Comfy Cloud with cloud-supported nodes only; local ComfyUI agent and custom-node support are still coming soon, and Comfy MCP is a separate external route.
    Comfy Agent launch announcementopens in a new tabDetails

  • Higgsfield releases Katana motion skill for Claude MCP

    What changed
    The Katana skill lets Claude trigger preset-driven video edits through Higgsfield MCP using up to five character, product, or clothing reference images and optional prompt adjustments.
    Why it matters
    Developers and designers already working in Claude can turn product or character reference photos into short motion clips without switching to a separate timeline editor.
    Limits to check
    Normal edits accept up to 5 reference images (40 is for community presets). Katana's entry documents preset video edits only; verify any alpha or format needs downstream.
    Higgsfield Creator Hub changelogopens in a new tabDetails

  • Langfuse expands experiment charts and baseline regression review

    What changed
    The Experiments view adds side-by-side metric charts, baseline summary statistics, per-item score threshold filtering, and keyboard-driven inline human scores, output corrections, and comments.
    Why it matters
    Teams evaluating a prompt, retrieval, or model change can spot where average gains hide a failing critical case and inspect the linked trace before shipping.
    Limits to check
    Comparisons require matching dataset and evaluator versions, missing scores are not passes, and picking a baseline in the UI does not configure an automated CI release gate.
    Langfuse experiment comparisons changelogopens in a new tabDetails

  • Superwhisper launches bot-free Meetings with local Mac models

    What changed
    Superwhisper 3.0+ on macOS records mic and system audio without a call bot, generates speaker-labeled transcripts and custom summaries, and can run offline with local transcription and LLM models.
    Why it matters
    With local transcription and LLM models selected, Mac teams can turn calls into local Markdown briefs without uploading audio or transcripts, or query saved notes from coding agents via MCP.
    Limits to check
    Requires Mac and Superwhisper 3.0+. Pages conflict on ongoing Free access vs. trial words (exhausted words stop recording); connected agents send what they read to their provider.
    Superwhisper Meetings launch postopens in a new tabDetails

  • Ollama adds local text-to-speech with Phonon on Apple Silicon

    What changed
    Ollama 0.40.11+ can stream 24 kHz WAV or PCM speech from text via the CLI or the OpenAI-compatible /v1/audio/speech endpoint using Gradium's MLX-accelerated Phonon models.
    Why it matters
    Developers already running Ollama on a Mac can pipe local model replies aloud or generate voiceover WAV files through standard OpenAI SDK audio calls.
    Limits to check
    TTS output only on Apple Silicon macOS (Ollama 0.40.11+; Windows/Linux planned). Base English Phonon is Apache 2.0, while 40M language variants are noncommercial CC BY-NC 4.0.
    Speech models in Ollama blog postopens in a new tabDetails

CodeSubmit team stack

The five tools our team uses

Codex, Claude Code, and Antigravity handle our repository work, while Hugging Face Hub and Ollama host and run our own models.

  1. Codex

    Our team uses Codex, especially with GPT-6 Astra. Start here when a repository change spans multiple files, tests, and architecture decisions, and choose Astra when complex work needs sustained reasoning and judgment.

    Configuration
    Select gpt-6-astra for demanding multi-step work; official docs recommend comparing gpt-6.1-sol on recurring tasks where usage cost matters.
    Tradeoffs & limits
    Model options and reasoning controls depend on your plan, client, and workspace settings, and API-key sessions bill at API rates without Codex Cloud or GitHub review.
    Codex model selection guideopens in a new tabDetails

  2. Claude Code

    Our team uses Claude Code. We recommend it for terminal-first repository edits, debugging failing test runs, and reviewing diffs in place when you want a direct loop over local files, Git history, and shell commands.

    Configuration
    Run in your project root with repo rules and local test commands, or pair with VS Code and JetBrains extensions for inline diff review.
    Tradeoffs & limits
    Excluded from Claude Free (requires Pro, Max, Team, Enterprise, Console, or a supported cloud provider), and third-party provider routing is limited to terminal and IDE clients.
    Claude Code quickstart guideopens in a new tabDetails

  3. Google Antigravity

    Dom started using agy at the Argon 4 Hackathon in San Francisco and likes it. With Gemini 4 Argon available in the desktop app, CLI, and IDE extensions, we recommend it for multi-step repository and agent tasks.

    Configuration
    Use the agy CLI, Antigravity App, or IDE extensions to select Gemini 4 Argon and adjust Low, Medium, or High thinking per turn.
    Tradeoffs & limits
    Gemini 4 Argon is not available in the standalone Antigravity IDE and requires paid Google AI Pro, Ultra, or eligible Enterprise access (or an eligible device Pro trial).
    Antigravity model availability docsopens in a new tabDetails

  4. Hugging Face Hub

    Our team uses Hugging Face for our own models. Use the Hub—not the separate Inference Providers router—when you need to inspect model cards and licenses, collaborate in Git-backed repos, or pull weights via hf CLI or Ollama.

    Configuration
    Run ollama run hf.co/{user}/{repo}:{quant} for GGUF repos; add ~/.ollama/id_ed25519.pub to Hugging Face settings for private GGUF access.
    Tradeoffs & limits
    The Hub hosts repos, not local inference; each checkpoint has its own license, format, and memory needs, and gated models require approved access (plus tokens for CLI or scripts).
    Hugging Face Hub documentationopens in a new tabDetails

  5. Ollama

    Our team uses Ollama to run our own models locally. It wraps local GGUF and Safetensors checkpoints behind a single CLI and OpenAI-compatible HTTP server so coding agents, eval runners, and apps can call local weights.

    Configuration
    Run Hub GGUF checkpoints directly with ollama run hf.co/... or import pre-quantized local GGUF files through a Modelfile with ollama create.
    Tradeoffs & limits
    Ollama does not quantize GGUF files during import, local throughput and context length depend on your machine's memory, and full cloud-catalog fallback requires paid credits.
    Use Ollama with Hugging Face Hub GGUFopens in a new tabDetails

Cross-tool workflows

Combining tools for a complete job

Six recipes from official docs showing how one tool's output feeds the next step, and where format or plan limits affect the handoff.

  • Run and evaluate Hugging Face weights locally

    Load a license-cleared GGUF model into Ollama and pass local Promptfoo checks before wiring the endpoint into your app.

    Starting input
    Target task prompt, Hub model card and license, compatible GGUF weights, and representative JSON/YAML test cases.
    Target output
    Pinned local Ollama model tag, Promptfoo assertion report, and an application module calling the local API.

    Tool handoffs

    1. Hugging Face Hub

      Verify the model card license, gated-access terms, and quantization files on Hugging Face Hub, then copy the repo path or download prepared GGUF weights.

    2. Ollama

      Load the model with ollama run hf.co/{user}/{repo}:{quant} (or ollama create from a Modelfile for pre-quantized local GGUF) and confirm it fits memory.

    3. Promptfoo

      Run promptfoo eval -j 1 against ollama:chat:<model> with task assertions, setting think: false when testing JSON schemas and keep_alive: 0 if RAM is tight.

    4. Codex

      Pass the pinned model tag, response schema, and failing edge cases to Codex to implement the application client and regression tests.

    Boundaries & checks
    Not every Hub repo ships GGUF or fits local RAM. Ollama does not quantize GGUF during import, and Promptfoo model-graded checks remain local only if you also set a local grader provider.
    Use Ollama with GGUF on Hugging Face Hubopens in a new tab

  • Animate a transparent image cutout from a voice brief

    Dictate the motion spec, strip a still image's background in Recraft, render alpha video in Remotion, and check playback in the browser.

    Starting input
    Spoken animation brief, a commercially licensed source image, and target light/dark UI backgrounds.
    Target output
    Downloaded alpha PNG/WebP cutout, transparent WebM (yuva420p) plus opaque fallback video, and browser QA screenshots.

    Tool handoffs

    1. Wispr Flow

      Dictate the animation timing, layout, and fallback requirements into your active Claude Code terminal or editor prompt as a plain-text brief.

    2. Recraft

      Call Recraft POST /v1/images/removeBackground on a rights-cleared raster asset and download the temporary PNG or lossless WebP cutout into the repo.

    3. Claude Code

      Have Claude Code wire the cutout into a React composition and app component with a transparent prop for alpha and opaque fallback renders.

    4. Remotion

      Render single-pass local WebM with --image-format=png --pixel-format=yuva420p --codec=vp8 (or ProRes 4444 for editors) plus an MP4 fallback.

    5. Playwright MCP

      Inspect the running localhost preview with Playwright MCP across light, dark, and reduced-motion states to confirm alpha compositing and playback.

    Boundaries & checks
    Recraft Free-plan images are public and noncommercial, API URLs expire in ~24 hours, Remotion preserves existing alpha rather than matting opaque video, and WebM alpha needs an opaque fallback.
    Remotion transparent videos guideopens in a new tab

  • Turn a PDF or Word spec into scoped code changes

    Parse the spec into auditable Markdown and JSON, locate the right symbols via LSP, and verify the code changes in a local browser.

    Starting input
    Specification PDF or DOCX, target repository with language-server support, and explicit acceptance criteria.
    Target output
    Source-linked Markdown/JSON spec, symbol-scoped repository diff, and verified test and browser results.

    Tool handoffs

    1. Docling

      Convert the specification PDF or DOCX with Docling into Markdown for reading and lossless Docling JSON to audit tables, formulas, and page references.

    2. Serena

      Initialize Serena MCP over the repository so the agent can look up symbol definitions, references, and file outlines instead of grepping raw text.

    3. Claude Code

      Supply the reviewed requirements to Claude Code to edit the target symbols and run unit tests against the specification's acceptance criteria.

    4. Playwright MCP

      Run Playwright MCP against the local preview to verify that the rendered UI states and form flows match each requirement in the spec.

    Boundaries & checks
    Check OCR and extracted tables against the source PDF before coding, treat document text as untrusted data rather than scope authority, and ensure Serena's language server is initialized.
    Docling supported formats documentationopens in a new tab

  • Test an AI feature locally and trace it in production

    Check your application handler with local Promptfoo assertions, trace live calls in Langfuse, and feed failing inputs back into your test suite.

    Starting input
    Application handler or prompt module, representative pass/fail examples, and model-provider credentials.
    Target output
    Versioned Promptfoo config and wrapper, OpenTelemetry trace spans in Langfuse, and updated regression cases.

    Tool handoffs

    1. Codex

      Use Codex to expose your application's actual prompt and tool-handling function through a thin Python provider wrapper (call_api) and add trace spans.

    2. Promptfoo

      Run npx promptfoo@latest eval against file://provider.py with deterministic schema and rubric assertions to catch regressions before deployment.

    3. Langfuse

      Instrument backend calls with the Langfuse SDK, inspect failing traces by version, and add redacted failure inputs to your Promptfoo regression suite.

    Boundaries & checks
    Promptfoo runs locally, but hosted targets or graders still receive data and bill tokens. Langfuse v5/v6 SDKs require Langfuse v4.59.0+, and turning traces into Promptfoo tests is a manual handoff.
    Langfuse observability SDK overviewopens in a new tab

  • Move a marimo notebook into a durable background job

    Validate data logic in a live marimo notebook, extract a tested Python function, and wrap each side effect in checkpointed Inngest steps.

    Starting input
    Sample dataset or schema, a running marimo notebook, and an event trigger or retry specification.
    Target output
    Validated marimo notebook, tested standalone transformation module, and an Inngest step.run() workflow.

    Tool handoffs

    1. marimo

      Start uvx marimo@latest edit notebook.py and connect your agent via marimo pair so it can inspect live dataframe schemas, test scratchpad logic, and run cells.

    2. Claude Code

      Have Claude Code extract the verified notebook cells into a pure repository function with unit tests and fixture checks that run without notebook state.

    3. Inngest

      Wrap each fetch, transform, and idempotent write in Inngest step.run() checkpoints so failed steps retry without rerunning completed earlier steps.

    Boundaries & checks
    marimo pair and Agent/Code modes can inspect live runtime state (and remote LLMs receive shared context). Inngest retries failed steps, so external writes inside step.run() must be idempotent.
    Pair with agents via marimo pairopens in a new tab

  • Matte a generated video clip into a transparent UI loader

    Generate a short motion clip in Higgsfield, strip the video background into alpha WebM in ComfyUI, and render browser-ready outputs in Remotion.

    Starting input
    Full-body key-pose image on a contrasting solid background, motion prompt, UI dimensions, and Comfy account credits.
    Target output
    ComfyUI SaveWEBM (vp9, yuva420p) alpha clip, Remotion transparent WebM + MP4 fallback, and Playwright browser checks.

    Tool handoffs

    1. Higgsfield

      Generate a 4s 720p MP4 with start=end pose, locked camera, and contrasting background; scrub to verify loop wrap (not guaranteed) and download within 7 days.

    2. ComfyUI

      In current/nightly ComfyUI (credits + API access), wire LoadVideo (MP4 <=60s) to Bria Remove Video Background (Transparent) (BriaTransparentVideoBackground).

    3. ComfyUI

      Feed Bria images+inverted mask to JoinImageWithAlpha (RGBA); wire RGBA + LoadVideo->GetVideoComponents fps to SaveWEBM (codec=vp9, yuva420p; av1 drops alpha).

    4. Remotion

      Pass VP9 WebM to <Video> from @remotion/media (no opaque background); render 1-pass local WebM (--image-format=png --pixel-format=yuva420p --codec=vp8) + MP4.

    5. Playwright MCP

      Open the localhost preview with Playwright MCP, emulate light/dark themes and reducedMotion, and check alpha compositing, loop wrap, and MP4 fallback playback.

    Boundaries & checks
    Untested recipe (no .riv/.lottie). Bria Partner Node uploads clips <=60s (needs current/nightly ComfyUI, positive credits, API access). AutoSprite is separate; Remotion keeps alpha (1-pass + MP4).
    ComfyUI Bria Partner Nodes tutorialopens in a new tab

Browse the collection

Coding harnesses

Terminal & desktop harnesses

Claude Code

Anthropic's coding agent that reads codebases, edits files, and runs shell commands across terminal, VS Code, JetBrains, desktop, and web clients.

claude.comDetails

Gemini CLI

Google's Apache-2.0 open-source terminal agent for querying and editing large codebases, generating apps from images or PDFs, and scripting MCP workflows.

Consumer free, Google AI Pro, and Ultra access ended June 18, 2026 (use Antigravity CLI); requires Code Assist Standard/Enterprise or paid Gemini API keys.

geminicli.comDetails

OpenCode

Open-source coding agent for running parallel project sessions across terminal, IDE, desktop, and web, with an embeddable server, SDK, and broad model support.

opencode.aiDetails

Pi

Extensible terminal coding agent with interactive, print, JSON event stream, RPC, and TypeScript SDK modes plus tree-structured session history.

pi.devDetails

Aider

Terminal pair-programming tool that maps your repository, edits code across 100+ languages, runs linters and tests automatically, and creates git commits.

aider.chatDetails

Cline

Open-source coding agent runtime with Plan and Act modes, workspace checkpoints, browser tools, a web Kanban worktree board, CLI, IDE extensions, and SDK.

cline.botDetails

Warp

Desktop terminal, CLI, and cloud coding agent platform that runs interactive local tasks with terminal and repository context or executes cloud runs through Oz.

warp.devDetails

Kilo

Coding agent and model gateway for VS Code, JetBrains, CLI, desktop, cloud, and mobile, supporting parallel isolated Git worktrees, MCP tools, and code review.

kilo.aiDetails

Plandex

Terminal coding agent that plans multi-file changes using Tree-sitter project maps, stages cumulative diffs in a review sandbox, and runs commands to debug builds.

Plandex Cloud began winding down on October 3, 2025 and no longer accepts new users; self-hosted local mode with your own API keys remains available.

github.comDetails

Wave Terminal

Desktop terminal with configurable Wave AI modes that connect local models, custom API endpoints, and cloud providers to interact with files and terminal widgets.

waveterm.devDetails

goose

Runs an open-source local agent via desktop app, CLI, API, or ACP editors to automate coding tasks with YAML recipes, parallel subagents, and MCP tools.

goose-docs.aiDetails

iTerm2

macOS terminal emulator with an optional AI Plugin and AI Chat that can explain output, watch sessions, or run commands under per-session permission controls.

iterm2.comDetails

AI code editors & IDE harnesses

Cursor

AI code editor and CLI with local and cloud agents, Bugbot pull-request reviews, Origin repository hosting, and TypeScript and Python SDKs.

cursor.comDetails

GitHub Copilot

GitHub's coding assistant and agent platform for inline completions, repository chat, CLI and editor agent mode, PR review, and assigned pull-request tasks.

github.comDetails

JetBrains Air

JetBrains' multi-agent workflow for running ACP coding agents, managing parallel sessions, and reviewing diffs in JetBrains IDEs, the terminal, or Air Teams.

jetbrains.comDetails

Kiro

Spec-driven IDE, CLI, and web environment that turns prompts into requirements and task plans, runs parallel agents, and verifies changes with property-based tests.

kiro.devDetails

Zed

Desktop code editor with native agentic editing, inline edit predictions, editable unified diff review, terminal-backed agent threads, and ACP and MCP integrations.

zed.devDetails

Amazon Q Developer

AWS coding and cloud operations assistant (formerly CodeWhisperer) for IDE completions, security scans, CLI chat, AWS Console guidance, and Java or .NET upgrades.

AWS will discontinue the IDE plugins on April 30, 2027 and directs IDE users to Kiro; CLI, AWS Console, and mobile surfaces continue.

aws.amazon.comDetails

Cody Enterprise

Sourcegraph's enterprise coding assistant for chat, auto-edits, and completions across IDEs, web, and CLI using multi-repository Sourcegraph Search context.

Sourcegraph discontinued Cody Free, Pro, and Enterprise Starter on July 23, 2025; standalone plans are no longer sold (requires Sourcegraph Enterprise).

sourcegraph.comDetails

Augment Code

Codebase-indexed agent platform combining VS Code and JetBrains extensions, the Auggie CLI, Context Engine MCP and SDK, and Cosmos scheduled or triggered workflows.

augmentcode.comDetails

TRAE

Downloadable coding environment and linked web workspace for code autocomplete and concurrent cloud agent tasks.

trae.aiDetails

Tabnine

Private, organization-aware enterprise coding agent and codebase chat for IDEs and CLI, backed by repository Context Engine indexing and admin governance.

docs.tabnine.comDetails

WPCode

WordPress admin snippet manager whose AI Snippet Generator drafts and refines custom PHP snippets while suggesting auto-insert locations and conditional logic.

wpcode.comDetails

aiXcoder

Private enterprise coding agent for VS Code, JetBrains, and CLI with Agent, Plan, and Ask modes, cross-file edits, MCP/terminal tools, and Enterprise Hub governance.

aixcoder.comDetails

Wing Pro

Python IDE (Wing 12 Pro) integrating Claude Code with a Tasks review queue, Code Actions, and MCP servers for Wing's static analysis, test runner, and debugger.

wingware.comDetails

Cloud & delegated workspaces

Codex

OpenAI's coding agent for local folders and reusable cloud environments across desktop, CLI, IDE, web, and mobile, with SDK automation and GitHub review.

learn.chatgpt.comDetails

Devin

Autonomous software engineering agent that writes, runs, and tests code in cloud or local sessions with an interactive browser, IDE, CLI handoff, and API.

devin.aiDetails

Devin Desktop

Desktop agent command center and full IDE (formerly Windsurf) for managing local and cloud coding agents, Kanban session boards, inline edits, and Tab completions.

devin.aiDetails

OpenHands

Coding agent platform combining a browser Agent Canvas, Python SDK, REST/WebSocket Agent Server, and repo automations triggered from GitHub, Slack, Jira, or Linear.

openhands.devDetails

Factory

Agentic engineering platform running Droid agents across desktop, CLI, web, and API to execute missions, review pull requests, and maintain AutoWiki repo docs.

factory.comDetails

Amp

Coding agent and development environment that runs each agent in a persistent hosted or self-hosted orb, letting tasks continue across desktop, terminal, and phone.

ampcode.comDetails

Roomote

Cloud or self-hosted coding agent by the Roo Code team that runs isolated dev sandboxes, cross-checks changes with a second model, and opens PRs with live preview URLs.

roomote.devDetails

AskCodi Desktop

Desktop multi-agent workspace where Codi plans tasks, delegates work across isolated Git worktrees, runs your local test suite, and stages results for review.

askcodi.comDetails

App & UI builders

Hosted full-stack app builders

Replit Agent

Browser-based coding agent that plans, builds, tests, and deploys web apps from chat with managed database and hosting integrations and Git checkpoints.

replit.comDetails

Lovable

Browser-based full-stack app builder that generates editable frontend and backend code with managed hosting, database, auth, payments, connectors, and Git sync.

lovable.devDetails

Bolt

Browser-based AI app builder that creates and refines websites and full-stack apps from prompts, Figma files, or GitHub repos with built-in Bolt Cloud hosting and auth.

bolt.newDetails

v0

Web app and prototype builder with visual Design Mode, design systems, backend and API integrations, GitHub sync, Vercel deployment, and an iOS companion app.

v0.appDetails

Base44

Browser app builder that generates hosted full-stack apps from prompts with managed database storage, authentication, payments, visual edits, and a testing agent.

base44.comDetails

Local & self-hosted app builders

bolt.diy

Open-source app-building workspace forked from Bolt.new that runs a WebContainer Node.js editor, terminal, and live preview with local models or your own API keys.

The public repository is in maintenance mode for bug fixes and dependency updates, with no new features planned.

github.comDetails

Dyad

Open-source desktop app builder that generates and visually edits web applications directly in local project folders, with Supabase, Neon, GitHub, and MCP support.

dyad.shDetails

UI prototyping & design handoff

Magic Patterns

Browser UI prototyping workspace that turns prompts and screenshots into interactive interface flows using your design system, with multiplayer canvas and MCP handoff.

magicpatterns.comDetails

ChatGPT

Interactive ChatGPT code blocks for editing code, previewing HTML, React, SVG, Mermaid, or Vega charts in chat, and running supported Python in a sandbox.

chatgpt.comDetails

Models & repositories

Coding & reasoning model weights

DeepSeek-V4.1-Flash

Multimodal model that reasons over text and images across a 1M-token context to generate code, structured JSON, and tool-call requests with adjustable thinking.

huggingface.coDetails

Devstral Small 2

Compact 24B vision-and-code model that inspects text and images across a 256K context to generate multi-file repository edits and tool-call requests.

Apache-2.0 weights remain downloadable; Mistral deprecated the hosted labs-devstral-small-2512 endpoint on February 27, 2026 in favor of Mistral Medium 3.5.

huggingface.coDetails

Gemma 4

Google open-weight family generating text, structured data, and function calls from text and images (plus audio on E2B, E4B, and 12B) across edge to 31B server sizes.

blog.googleDetails

GLM-5.3

Text-only reasoning model that generates code, plans, and tool-call requests with always-on adjustable thinking and up to 1M API context (128K max output).

huggingface.coDetails

gpt-oss

OpenAI open-weight reasoning family in 120B server and 20B local tiers that generates adjustable reasoning traces, structured data, and formatted tool-call requests.

openai.comDetails

Kimi K3

Multimodal reasoning model that analyzes text, images, and video across a 1M-token context with always-on thinking to emit multi-turn tool calls and code edits.

huggingface.coDetails

Leanstral 1.5

Formal-verification model that accepts text and images to generate and repair Lean 4 proofs and formalize software specifications for proof-checking harnesses.

Apache-2.0 weights remain available to download; Mistral marked the hosted labs-leanstral-1-5 endpoint deprecated on September 29, 2026.

huggingface.coDetails

Qwen3-Coder-Next

Non-thinking code generation model with a 262K native context that emits repository edits, fill-in-the-middle completions, and tool-call requests for agent harnesses.

qwen.aiDetails

Speech, retrieval, safety & vision weights

Chatterbox Turbo

Resemble AI's MIT-licensed 350M English text-to-speech model with a single-step mel decoder, reference-audio voice cloning, and paralinguistic tags ([laugh], [cough]).

huggingface.coDetails

Qwen3-TTS

Apache-2.0 open-weight 0.6B and 1.7B speech models covering 10 languages across dedicated Base (voice cloning), CustomVoice (9 presets), and VoiceDesign checkpoints.

github.comDetails

Model repositories & weight hubs

Hugging Face Hub

Git-based repository platform hosting models, datasets, and Spaces, with model cards, version history, and weight downloads via libraries, the hf CLI, or Git.

huggingface.coDetails

Inference APIs & local runtimes

First-party foundation model APIs

OpenAI API

Developer API platform and SDKs exposing GPT-6.1 Sol, GPT-6 Astra, GPT-6 Luna, and GPT-5.5 models through Responses, Chat Completions, Decisions, and Agents APIs.

openai.comDetails

Claude Platform

Developer console and API for calling Fable 5.1, Opus 5.5, Sonnet 5.5, and Haiku 5.5 with prompt caching, batch processing, and cloud provider integrations.

claude.comDetails

Gemini API

Google Gen AI SDKs, Interactions API, and AI Studio for integrating Gemini (including Gemini 4 Argon), Veo, and Nano Banana text, image, video, and audio models.

ai.google.devDetails

Mistral Studio

Developer platform and APIs for building and evaluating stateful agent workflows on Mistral Large 4, Codestral, OCR, embedding, and Voxtral audio models.

mistral.aiDetails

DeepSeek API

OpenAI- and Anthropic-compatible API serving DeepSeek V4.1 Flash and V4 Pro models with thinking modes, tool calling, JSON output, and model-dependent vision or FIM.

Older Flash and vision model implementations are retired; their legacy slugs now route to V4.1-Flash-1017 and bill at replacement Flash rates.

api-docs.deepseek.comDetails

Grok API

Developer REST and Responses APIs for Grok 4.7 and Grok Build models, plus Imagine image/video and voice services with tool calling, remote MCP, and web/X search.

On November 2, 2026, grok-imagine-image-quality retires and its slug redirects to image-2.0 at low quality with replacement pricing.

docs.x.aiDetails

Kimi API

Moonshot's developer API serving Kimi K3.1 Pro, K3, K2.7 Code, and K2.6 for coding, reasoning, and model-dependent vision via OpenAI- and Anthropic-compatible APIs.

K2 series models close to new users on October 31, 2026, and all K2 series models are scheduled to retire by the end of December 2026.

platform.kimi.aiDetails

Z.ai

Developer API platform serving GLM-5.3, GLM-5.3-Flash, and GLM-5.1 models through REST, Python and Java SDKs, and OpenAI-compatible client endpoints.

docs.z.aiDetails

Multi-provider routers & GPU clouds

OpenRouter

Unified OpenAI-compatible API router for calling hosted models across multiple providers with streaming, provider selection, and automatic fallback routing.

openrouter.aiDetails

Replicate

Cloud API platform for running public language, image, video, and audio models or packaging and deploying private custom models and dedicated instances with Cog.

replicate.comDetails

Amazon Bedrock

Managed AWS API platform for building generative AI applications and agents with Amazon and third-party foundation models, Knowledge Bases, customization, and Guardrails.

aws.amazon.comDetails

Baseten

Managed inference and training cloud offering token-billed Model APIs, Frontier Gateway routing, and custom Python, Truss, or Docker deployments on dedicated GPUs.

baseten.coDetails

fal.ai

Generative media API platform hosting image, video, audio, 3D, and background-removal endpoints alongside serverless GPU compute for custom model deployments.

fal.aiDetails

Fireworks AI

Inference and training platform offering OpenAI-compatible serverless APIs, dedicated GPU deployments, and fine-tuning across text, vision, audio, and image models.

fireworks.aiDetails

Groq

Hosted GroqCloud inference platform exposing OpenAI-compatible and Responses APIs for text, reasoning, tool-calling, structured-output, and speech models.

console.groq.comDetails

Modal

Serverless cloud platform for running containerized Python functions, GPU inference and training jobs, isolated agent Sandboxes, and token-billed Shared Endpoints.

modal.comDetails

Together AI

Cloud platform providing OpenAI-compatible serverless APIs, dedicated model and container endpoints, fine-tuning, and GPU clusters across open-weight models.

together.aiDetails

AI gateways & proxy routers

LiteLLM

Python SDK and self-hosted AI gateway that routes OpenAI-format requests across model providers with virtual keys, rate limits, budgets, and team cost tracking.

docs.litellm.aiDetails

Helicone

Routes LLM calls through an AI gateway that logs requests, sessions, costs, and user metrics while providing caching, retries, rate limits, and prompt tools.

helicone.aiDetails

Portkey

Unifies model-provider APIs behind a gateway with routing, retries, fallbacks, caching, prompt management, provider-key controls, and request observability.

portkey.aiDetails

Cloudflare AI Gateway

Hosted proxy and Cloudflare Workers integration that adds caching, rate limiting, routing, analytics, and optional unified billing in front of model API calls.

Gateways created on or after September 24, 2026 follow Workers Logs pricing and retention, and the legacy OpenAI-compatible Unified API is deprecated.

developers.cloudflare.comDetails

Local runtimes & serving engines

Ollama

CLI and API runtime for running open models on local hardware or through Ollama's hosted cloud catalog via Ollama, OpenAI, and Anthropic-compatible endpoints.

ollama.comDetails

llama.cpp

C/C++ inference runtime and library for running quantized GGUF language and vision models via CLI, web UI, or an OpenAI-compatible HTTP server.

github.comDetails

LM Studio

Local model runtime with a desktop GUI, `lms` CLI, and headless `llmster` daemon that serves downloaded models over OpenAI-compatible local APIs.

lmstudio.aiDetails

MLX LM

Apple MLX-based Python library and CLI for local text generation, model quantization, and LoRA or full fine-tuning on Apple silicon with Hugging Face Hub support.

github.comDetails

vLLM

Open-source inference engine and API server with PagedAttention, continuous batching, prefix caching, and OpenAI, Anthropic Messages, and gRPC endpoints.

github.comDetails

Lemonade

Local AI server, desktop app, and CLI serving GGUF, FLM, and ONNX chat, speech, and image models over OpenAI-, Anthropic-, and Ollama-compatible HTTP APIs.

lemonade-server.aiDetails

Agent tooling & MCP

Agent & application SDKs

Claude Agent SDK

Python and TypeScript SDKs exposing the Claude Code agent loop, context management, file/shell/search tools, subagents, sessions, hooks, MCP, and skill loading.

code.claude.comDetails

LangGraph

Low-level orchestration runtime for stateful, long-running agent graphs with durable execution, checkpoint persistence, streaming, and human-in-the-loop controls.

docs.langchain.comDetails

Pydantic AI

Python agent framework that enforces Pydantic schema validation on model outputs, tool arguments, and injected dependencies across multiple LLM providers.

pydantic.devDetails

Vercel AI SDK

TypeScript library pairing multi-provider text, structured-object, and tool-calling APIs with streaming UI hooks for React, Next.js, Svelte, Vue, and Angular.

ai-sdk.devDetails

MCP SDKs, servers & SKILL.md tools

FastMCP

Python framework by Prefect for building MCP servers, clients, and interactive tool apps with automatic schema generation, validation, auth, and transport handling.

gofastmcp.comDetails

MCP Inspector

Developer testing suite providing web UI, CLI, TUI, and experimental mcpdo daemon tools to inspect MCP servers, exercise tools, prompts, and resources, and debug auth.

github.comDetails

MCP TypeScript SDK

TypeScript client and server libraries implementing the 2026-07-28 MCP specification over stdio and Streamable HTTP across Node.js, Bun, and Deno.

github.comDetails

MCP registries & server directories

Glama MCP Registry

Searchable third-party MCP server directory with filters for tools, prompts, resources, Python/TypeScript runtimes, local or remote deployment, and maintenance grades.

glama.aiDetails

PulseMCP

Web directory of MCP servers and clients paired with implementation guides, integration use cases, and the Agentic Loop ecosystem newsletter.

The site remains live for browsing, but new server submissions and listing edits are paused during an ingestion rework with no announced reopening date.

pulsemcp.comDetails

Smithery MCP Registry

Public, non-proxied web directory for looking up MCP servers and their direct vendor connection URLs, maintained by Arcade.

On October 31, 2026, Smithery shuts down hosted connection management, accounts, API/Connect/Toolbox, and proxy URLs; the public registry stays online.

smithery.aiDetails

Symbol & library documentation context

Serena

Local MCP server and coding toolkit that gives AI agents symbol-level code retrieval and editing through language servers or a JetBrains plugin backend.

github.comDetails

Context7

Documentation retrieval service, CLI skill installer (ctx7), and MCP server that injects version-specific library docs and code examples into coding agents.

context7.comDetails

Jedi

Deterministic Python static-analysis library providing autocompletion, type inference, goto, reference lookup, code search, and refactoring APIs for Python 3.10+.

jedi.readthedocs.ioDetails

Browser automation & cloud sessions

Playwright MCP

Microsoft MCP server that lets coding agents drive Chrome, Firefox, WebKit, and Edge using Playwright's accessibility tree, page actions, and screenshots.

github.comDetails

agent-browser

Native Rust CLI and daemon that gives shell-capable coding agents browser navigation, accessibility snapshots with element refs, screenshots, tracing, and diffs.

agent-browser.devDetails

Browser Use

Open-source Python library, CLI, and hosted API for running browser agents over local or remote CDP browsers with structured outputs and custom tools.

browser-use.comDetails

Stagehand

Browser automation SDK for TypeScript, Python, and Go that mixes direct Chromium CDP page control with natural-language act, observe, and schema-validated extract.

stagehand.devDetails

Browserbase

Provisions, controls, and observes headless browser sessions in the cloud, pairing managed browser infrastructure with web search/fetch APIs and sandboxes.

browserbase.comDetails

Steel

Provides an API for launching and controlling remote browser session fleets for AI agents, with managed cloud and open-source self-hosted deployments.

steel.devDetails

Review, testing & security

Pull-request review & stacked diffs

CodeRabbit

Automated code review and change-management platform for PRs, IDEs, and CLIs featuring inline fix suggestions, PR queue triage, blast-radius audits, and security scans.

coderabbit.aiDetails

Qodo

Multi-agent code review and quality platform offering PR standards enforcement, cross-repo impact mapping, IDE plugins, and a CLI quality toolbox for coding agents.

qodo.aiDetails

Graphite

Creates and merges stacked GitHub pull requests with AI code reviews, PR chat, a unified inbox, and stack-aware merge queues.

graphite.comDetails

Greptile

Indexes repository graphs to post context-aware pull request reviews, suggested fixes, and custom rule checks in web or CLI workflows.

greptile.comDetails

Unit & end-to-end test automation

Diffblue Agents

Local CLI orchestrator that drives Claude Code, Copilot CLI, Gemini CLI, or OpenCode in temporary Git worktrees to generate and validate Java and Python unit tests.

diffblue.comDetails

mabl

Generates, runs, and repairs web, native mobile, and API tests while exporting screenshots, HAR files, and console logs through its CLI and MCP.

mabl.comDetails

Meticulous

Records user and developer sessions, then replays selected flows against base and pull request builds to flag visual and behavioral regressions.

meticulous.aiDetails

Momentic

Runs repository-backed YAML tests with AI assertions and exploratory Mo sessions across web, iOS, and Android, returning traces and repair proposals.

momentic.aiDetails

QA Wolf

Generates customer-owned Playwright and Appium end-to-end tests from natural language and executes them on managed parallel infrastructure.

qawolf.comDetails

Shortest

Executes natural-language browser and API tests on top of Playwright using Anthropic Claude API calls, TypeScript callbacks, and lifecycle hooks.

github.comDetails

Playwright

Deterministic end-to-end test runner and browser automation library for Chromium, Firefox, and WebKit with isolated contexts, Codegen recording, and Trace Viewer.

playwright.devDetails

Source code & dependency security

Semgrep Multimodal

Combines static analysis with AI reasoning on the Semgrep AppSec Platform to detect issues, triage alerts, and post fix guidance on pull requests.

semgrep.devDetails

Snyk DeepCode AI

Powers Snyk Code static application security testing to trace data flows, prioritize vulnerabilities, and suggest fixes across IDEs, CLIs, and repos.

snyk.ioDetails

Socket

Analyzes open-source dependencies for supply-chain risks and vulnerabilities across GitHub pull requests, package-manager installs, VS Code, and CLI runs.

socket.devDetails

Media & UI assets

Raster & vector image generation

Nano Banana

Gemini image generation and conversational editing model family producing PNG and JPEG visuals from text, image, and selected video references with Search grounding.

ai.google.devDetails

FLUX

Black Forest Labs models for FLUX.2 image generation and multi-reference editing up to 4MP, plus FLUX 3 HD/FHD video generation up to 20 seconds with optional audio.

bfl.aiDetails

Recraft

Design workspace and REST API that generate editable SVG vectors and 1K/2K raster images, with built-in raster-to-SVG conversion and WebP/PNG background removal.

recraft.aiDetails

Ideogram

Generates text-accurate raster images, UI cutouts, and native transparent PNGs with Ideogram 4.5 and 4.0 Transparent models in the web app or v2 REST API.

ideogram.aiDetails

Seedream 5.0 Pro

ByteDance's hosted text-to-image and reference-guided editing model on Replicate, generating 1K or 2K PNG and JPEG images from prompts or up to 10 reference images.

replicate.comDetails

Wan 2.7 Image Pro

Alibaba Wan's hosted image model on Replicate for text-to-image up to 4K, multi-image editing with up to 9 reference images, and coherent sets up to 12 images.

replicate.comDetails

Google Upscaler

Hosted image upscaling endpoint on Replicate (google/upscaler) that enlarges existing or generated raster images by 2x or 4x through the web playground or HTTP API.

replicate.comDetails

Video generation & clip editing

Runway

Web creative workspace and Runway Dev API for generating and editing video, images, and audio, with web Gen-4.5 tools, 4K upscaling, and MP4, ProRes, or PNG exports.

runway.comDetails

Gemini video generation

Video generation and multi-turn conversational editing in the Gemini Interactions API via Gemini Omni Flash, outputting MP4 clips with audio up to upscaled 4K.

Veo 3.1 Preview endpoints in the Gemini API shut down on October 22, 2026; Google directs callers to Gemini Omni Flash or GA Veo on Enterprise.

ai.google.devDetails

Luma Dream Machine

Pairs the consumer Dream Machine web app with the async Luma Agents REST API for Ray 3.2 video generation, looping, and reframing (MP4 up to 1080p) and uni-1 images.

lumalabs.aiDetails

Kling Video 3.0

Hosted Kling 3.0 video generator on Replicate for 3–15s text-to-video and start/end-frame animation at 720p or 1080p, with optional native audio and 6-shot prompts.

replicate.comDetails

Kling Video 3.0 Omni

Multimodal Kling 3.0 Omni endpoint on Replicate combining 720p/1080p video generation (up to 15s), multi-image character/style references, and reference-video editing.

replicate.comDetails

LTX 2.3 Pro

Hosted Lightricks video model on Replicate for text/image-to-video up to 4K (6/8/10s, 24–50 FPS) plus 1080p audio-to-video, clip retake, and extend operations.

replicate.comDetails

P-Video

Pruna AI's hosted text-, image-, and audio-to-video model on Replicate with a fast draft preview mode, native or imported audio, and 720p/1080p output at 24 or 48 FPS.

replicate.comDetails

Programmatic video, UI motion & 3D

Remotion

Programmatic React video framework and renderer that composites parameterized motion graphics and encodes transparent WebM (yuva420p) or ProRes 4444/4444-xq video.

remotion.devDetails

LottieFiles

Creates and edits scalable vector animations with Prompt to Lottie, Motion Copilot, and MCP tools, exporting .lottie, Lottie JSON, SMIL SVG, or transparent WebM/MOV.

lottiefiles.comDetails

Rive

Builds state-machine-driven .riv animations via the Rive CLI or visual editor for web, mobile, desktop, and game runtimes, plus cloud-rendered video and sequences.

Rive's pricing page lists Free .riv exports with a splash screen, while the runtime-export docs still state that .riv export requires a paid plan.

rive.appDetails

Spline

Creates interactive 3D scenes in the browser with AI generation and Spline MCP control, exporting React/native embeds, 3D meshes, MP4/WebM video, and PNG/JPG sequences.

spline.designDetails

Node-graph & sprite asset pipelines

ComfyUI

Builds reusable node-graph workflows for image, video, audio, 3D, and text generation, then executes them locally, on Comfy Cloud, or through HTTP and WebSocket APIs.

comfy.orgDetails

Higgsfield

Generates images and motion clips via web tools or an async API with webhooks, plus documented 2D spritesheet workflows for UI loaders and game assets.

higgsfield.aiDetails

Music & soundtrack generation

Lyria

Google's music generation model family in the Gemini Interactions API producing 44.1 kHz stereo MP3 tracks with vocals, timed lyrics, and instrumentals.

ai.google.devDetails

Voice & speech

Voice dictation & meeting notes

Superwhisper

Dictates speech into editors and terminals using local Whisper models or cloud AI modes, with a macOS CLI, stdio MCP server, and coding-agent plugins.

superwhisper.comDetails

Wispr Flow

Dictates prompts, code comments, and PR notes into active apps on Mac, Windows, iOS, and Android, with IDE identifier recognition and voice snippets.

wisprflow.aiDetails

Speech-to-text & text-to-speech APIs

ElevenLabs

Audio platform with REST, TypeScript, and Python SDKs for speech synthesis, batch and real-time transcription, voice agents, dubbing, sound effects, and music.

elevenlabs.ioDetails

Cartesia

Streams synthesized speech with Sonic TTS, transcribes audio with Ink STT, and runs managed voice agents over HTTP, SSE, and WebSocket APIs.

cartesia.aiDetails

Deepgram

Provides REST and WebSocket APIs for speech-to-text (including Nova-3 and Deepgram Flux), text-to-speech, and conversational voice agents.

deepgram.comDetails

Inworld TTS

Hosted text-to-speech platform serving Realtime TTS-2 and TTS-2 Flash via streaming and batch APIs, SDKs, and Playground across 200+ locales with instant voice cloning.

docs.inworld.aiDetails

Realtime voice-agent frameworks

LiveKit Agents

Open-source Python and Node.js framework that bridges WebRTC or telephony users to STT, LLM, and TTS pipelines with turn detection and tool calls.

docs.livekit.ioDetails

Local & self-hosted Whisper runtimes

Whisper-Flow

Self-hosted Python framework and WebSocket server (whisperflow) that streams 16 kHz mono PCM audio into partial and final JSON Whisper transcripts.

github.comDetails

whisper.cpp

C/C++ Whisper runtime for local speech-to-text across CPU and GPU backends, with a C API, HTTP server example, and CLI exporting TXT, VTT, SRT, CSV, LRC, or JSON.

github.comDetails

Data & workflow automation

Web search, crawling & document parsing

Marker

Document ingestion toolkit that converts files into structured Markdown, JSON, chunks, or HTML across CLI, API, and self-hosted runs with optional LLM processing.

github.comDetails

Apify

Runs hosted web-scraping and automation Actors, exposing collected datasets to coding agents, RAG pipelines, and vector databases through MCP and SDK integrations.

apify.comDetails

Crawl4AI

Crawls web pages via Python browser automation to emit clean Markdown and structured data using deterministic CSS/XPath rules or configured LLM extraction.

docs.crawl4ai.comDetails

Exa

Provides web search and content retrieval APIs that return page text and highlights for coding agents, chatbots, and research pipelines.

exa.aiDetails

Firecrawl

Searches, crawls, and scrapes web pages into clean Markdown or structured data, with support for interactive browser sessions.

firecrawl.devDetails

Jina Reader

Converts web URLs into clean, LLM-ready text through an HTTP API, with documented search and MCP interfaces for agent workflows.

jina.aiDetails

Tavily

Provides real-time web search, page extraction, site crawling, mapping, and research endpoints for agent workflows via direct API or MCP connections.

tavily.comDetails

Reactive notebooks & SQL data layers

marimo

Reactive Python notebook environment with AI-assisted cell refactoring, full-notebook generation, and variable-aware context.

marimo.ioDetails

MotherDuck

Managed DuckDB cloud analytics platform that exposes separate MCP tools for SQL reads and writes so agents can query and update analytical datasets.

motherduck.comDetails

Supabase AI tools

Combines a browser dashboard assistant for SQL and schema work with an MCP server, agent skills, and plugins that connect coding agents to Supabase projects.

supabase.comDetails

Git-synced documentation publishing

Mintlify

Hosts Git-backed developer documentation with an embedded answer assistant that cites sources and agents that help draft and maintain docs.

mintlify.comDetails

Durable jobs, workflows & SaaS auth

Dagger

Runs programmable container workflows that expose selected environment and container objects as callable tools for LLM agents.

dagger.ioDetails

Hatchet

Executes code-defined tasks and durable workflows with queuing, automatic retries, agent checkpoints, monitoring, and run replays.

hatchet.runDetails

Inngest

Runs event-driven functions that checkpoint state, wait on external events, and coordinate multi-step agent retries and flow control.

inngest.comDetails

n8n

Builds visual, node-based automations that connect LLM agents, tool calls, memory stores, and external application APIs.

n8n.ioDetails

Pipedream Connect

Provides managed OAuth, credential refresh, prebuilt actions, a raw API proxy, and remote MCP endpoints so agents can act on user-connected SaaS accounts.

Pipedream Workflows and String are in maintenance mode and shut down on March 31, 2027; Pipedream Connect is unaffected.

pipedream.comDetails

Trigger.dev

Runs long-lived background tasks and durable AI agent sessions with retries, orchestration, and realtime task visibility.

trigger.devDetails

Windmill

Orchestrates AI agent steps alongside scripts, MCP tools, database queries, and approval or error branches inside workflow pipelines.

windmill.devDetails

Evals & observability

Evaluation runners & prompt optimizers

DSPy

Declarative Python framework for building modular language-model pipelines and running algorithms that compile and optimize prompts or model weights against metrics.

dspy.aiDetails

DeepEval

Runs Pytest-style LLM evaluations for outputs, agent trajectories, and traced components, with a skill workflow for coding agents to iterate on failing tests.

deepeval.comDetails

Promptfoo

Evaluates and red-teams LLM apps via a local CLI, library, or CI workflow, running declarative comparisons across prompts, models, and RAG pipelines.

promptfoo.devDetails

Ragas

Evaluates RAG retrieval and answer quality, agent tool calls, and text-to-SQL workflows, with synthetic test-data generation and dataset experiments.

docs.ragas.ioDetails

Tracing, datasets & experiment platforms

Arize Phoenix

Provides open-source local and self-hosted AI observability with OpenTelemetry/OpenInference tracing, versioned datasets, LLM evaluations, and prompt experiments.

arize.comDetails

Braintrust

Connects AI tracing, human annotation, evaluation datasets, and experiments across a web UI, bt CLI, MCP server, and the Loop trace-analysis agent.

braintrust.devDetails

Langfuse

Traces model calls and agent workflows, tracks cost and latency, versions prompts, and runs dataset evaluations via web UI, API, CLI, or MCP server.

langfuse.comDetails

LangSmith

Captures execution traces across supported AI frameworks and SDKs, builds evaluation datasets from runs, monitors quality, and collects human feedback.

langchain.comDetails

W&B Weave

Traces LLM application and agent calls, versions instrumented code, and evaluates outputs using LLM judges or custom scoring functions.

W&B Weave docs route to CoreWeave: Agent Lens is the agent-tracing successor sharing trace ingestion, while Weave function and evaluation workflows remain.

docs.coreweave.comDetails

Incident & telemetry debugging agents

HolmesGPT

Open-source SRE agent that investigates production incidents, telemetry signals, and deployment issues from a standalone CLI or self-hosted operator.

holmesgpt.devDetails

Sentry Seer

Analyzes production issues, plans fixes, and reviews code by combining Sentry application telemetry with connected source repositories.

sentry.ioDetails

How this collection is maintained

We mark the five tools our team uses, while all profile facts, release notes, and workflow recipes are compiled from official docs and pricing pages rather than controlled testing. Check the linked sources for current plans and limits.

Original historical article
Suggest a tool or update