Radar
New AI Tools 2026 (June Week 3): Builder Lane Focus
New AI tools 2026 (June week 3): builder lane focus with Devin Desktop, Codex Desktop, Copilot Workspace, OpenCode, and more. Updated June 2026.
June Week 3, 2026 is the builder lane focus for new ai tools 2026. We scanned Product Hunt, AIxploria, and TAAFT on June 5, 2026. This week is heavy on builder-lane tools because early June 2026 saw major releases from OpenAI (Codex Desktop), GitHub (Copilot Workspace), and Cognition (Devin Desktop rebrand). We also caught new productivity and design tools. We dropped Perplexity Pages and ChatGPT Agent mode as they are not standalone product slots worth a radar entry. Full agent and model context still lives in Manus AI Review (2026) and Latest AI Models Compared (2026).

Changelog
- 2026-06-05: Week 3 update. Scanned Product Hunt, AIxploria, TAAFT on June 5, 2026. Added Devin Desktop (rebranded from Windsurf), Codex Desktop, Copilot Workspace, OpenCode, Mina, Wispr Flow, Astra, Lumo Studios, and Ideogram 4.0. Dropped Perplexity Pages and ChatGPT Agent mode as not relevant.
Method in one paragraph
Week 3 uses the same gate as other radar posts (see Editorial Policy). The extra step is a pricing worksheet: free tier cap, paid entry price, export paywall, and whether failed jobs still burn credits. We publish numbers only when we saw them on a logged-in account or a official pricing page the same week.
Slides
1. Dokie AI
Verdict: Watch (slides; aligns with full review)
Dokie AI was Watch in our full review after the June 2 fact-check. Week 3 adds pricing proof, not a verdict upgrade. We ran two decks: a ten-slide QBR and a six-slide training deck. Generation worked on the free credit pool; PPTX export is listed as a paid-plan perk on dokie.ai/pricing (June 2026).
Free tier signal: Starter includes welcome + daily credits (~100 credits per new deck per FAQ). Heavy iteration burns credits before you reach export.
Paid signal: Plus/Pro tiers add export options and custom templates. Screenshot your checkout screen. Do not trust third-party listicles for dollar amounts.
Best for: Teams that must deliver .pptx files, not only web scroll decks.
Skip if: You want zero Office dependency. Try Gamma below.
- Lane: Creators / slides
- Competes with: SlideAI, Gamma, Copilot in PowerPoint
- Cross-read: Week 1 Dokie section for workflow prompts (see June Week 1 radar)
2. Gamma
Verdict: Use (visual web decks)
Gamma remains the default visual deck tool when you want web-native presentation flow and built-in themes. June 2026 pricing still uses AI credits for generation and some exports.
Free tier check (June 5): Free accounts can create decks with a monthly credit pool. PDF export worked on our test account. PPTX export rules vary by plan on gamma.app/pricing; verify live before you promise a client a .pptx file. That is the main gap vs Dokie Plus for Office-first teams.
Best for: Startup pitches, internal updates where you share a link, not a file attachment.
Skip if: Your client demands PowerPoint and will not accept a Gamma link.
- Lane: Creators / slides
- Competes with: Dokie, SlideAI
- Disclosure: We did not build Gamma. Compare fairly with SlideAI review for credit math
Builders
3. Devin Desktop (formerly Windsurf)
Verdict: Use (builders, IDE)
Devin Desktop is the new name for Windsurf as of June 2, 2026. It is a Codeium-backed AI IDE with agent flows and inline edits. The rebrand added the Agent Command Center (Spaces, Kanban view, multi-agent management) as the primary view. June 2026 autocomplete shows devin desktop vs cursor 2026 as a top builder query. That alone justifies a Use radar slot ahead of our full comparison post.
Pricing signal: Free tier with limited agent credits. Pro tiers unlock more premium model calls (verify live on devin.ai/pricing). Teams should track premium model surcharges separately from base subscription. Pricing is unchanged from the Windsurf era.
What we tested: One React bugfix across three files. Agent proposed a patch set. We ran tests locally before merge. Latency was acceptable on Pro trial.
Best for: Developers who want agentic edits inside a VS Code-like UI, plus multi-agent management through the Agent Command Center.
Skip if: You are happy with Cursor and do not want to relearn keybindings.
- Lane: Builders
- Competes with: Cursor, GitHub Copilot
- Compare: Devin Desktop vs Cursor (2026) already published
4. Cursor (June 2026 build)
Verdict: Use (builders, IDE)
Cursor is the benchmark IDE for AI pair programming in 2026. Week 3 is not a full Cursor review. It is a pricing refresh: Hobby tier still exists with limited agent requests. Pro still centers on frontier model access with monthly caps on some agent modes (verify live).
What changed in June: More agent-style background tasks in marketing pages. Check whether your plan bills per request or per token. Mixed billing confuses teams.
Best for: Daily coding with chat + composer + repo-wide context.
Skip if: You only write SQL in a locked-down enterprise IDE with no extensions.
- Lane: Builders
- Pair with: Devin Desktop section for compare context
- Models: Route hard bugs to GPT-5.5 or Claude per models hub
5. Kimi PPT
Verdict: Watch
Kimi’s presentation mode keeps showing up in Chinese and global directories with strong slide output demos. Search demand in English is still weaker than dokie ai ppt or slide ai ppt. We stay Watch until we finish English UI export tests and pricing in USD.
Free tier signal: Chat-side free usage is generous. PPT-specific export limits are unclear without a logged-in test in your locale.
Best for: Bilingual teams already using Kimi for long documents.
Skip if: You need US-friendly support docs and predictable PPTX export today.
- Lane: Creators / slides
- Month 2: Best AI Presentation Maker comparison cluster
6. OpenAI Codex Desktop
Verdict: Use (builders, IDE agent)
OpenAI Codex Desktop is the new standalone coding app from OpenAI. It pairs GPT-5.5 with a visual file explorer, diff viewer, and terminal panel in one window. You describe what you want built. Codex plans the changes, edits multiple files, runs tests, and shows you a diff before you accept. It is not a CLI tool. It is a full desktop experience for developers who want OpenAI’s latest agent in a polished UI.
Pricing signal: Free tier with daily limits as of June 2026. Pro tiers unlock more agent runs and larger context windows. API usage may bill separately on top of subscription. Verify live on openai.com/codex before you commit.
What we tested: One Python API scaffold with FastAPI and Pydantic models. Codex proposed the folder structure, generated main.py and models.py, ran uvicorn locally, and fixed a type error on the first retry. Total wall time: about 7 minutes.
Best for: Developers who want OpenAI’s agent stack in a dedicated app without configuring an IDE plugin.
Skip if: You need MCP tool hooks, deep GitHub integration, or you already standardized on Cursor / Devin Desktop.
- Lane: Builders
- Competes with: Cursor, Devin Desktop, Claude Code
- Hub: GPT-5.5 rows on models hub
7. GitHub Copilot Workspace
Verdict: Use (builders, GitHub-native agent)
GitHub Copilot Workspace is the agent layer inside GitHub that turns issues into planned, coded, and tested pull requests. June 2026 builds add multi-file awareness and better test generation. It lives inside GitHub, not a separate IDE, which matters for teams that want agent features without switching editors.
Pricing signal: Copilot Individual ($10/mo), Business ($19/user/mo), and Enterprise ($39/user/mo) tiers include Workspace features with varying limits. Free tier is limited to open-source public repos. Verify live on github.com/pricing.
What we tested: One issue-to-PR flow on a public repo. Workspace read the issue, proposed a plan, generated the code, and opened a draft PR. Human review still required before merge.
Best for: Teams already on GitHub who want agentic PR generation without buying a second IDE.
Skip if: You need deep IDE integration, MCP servers, or you do not use GitHub for repo hosting.
- Lane: Builders
- Competes with: Cursor cloud agents, Devin Desktop Cascade, Replit Agent
- Pair with: Copilot inline completion for the daily editing layer
Productivity
8. Mina Meeting Assistant
Verdict: Use
Mina launched on Product Hunt in early June 2026 as an AI teammate that joins your calls, transcribes, and executes tasks during the meeting. It is not a passive notetaker. It answers questions, drafts follow-up emails, and updates CRM fields while the call is still live.
Pricing signal: Free tier for personal use with limited meeting minutes. Pro tiers unlock CRM integrations and longer sessions. Verify live on mina.ai/pricing.
Best for: Sales reps and founders who live on Zoom and want async task execution without hiring a human assistant.
Skip if: Your company blocks third-party meeting bots for compliance reasons.
- Lane: Productivity / Agents
- Competes with: Otter, Fireflies, Granola
- PH rank: Top 5 daily launch June 1, 2026
9. Wispr Flow
Verdict: Use
Wispr Flow is an AI dictation tool that claims 4x faster input than typing. It runs locally on macOS and works inside any text field. June 2026 feedback from early users says accuracy is strong for technical vocabulary, which is where most dictation tools fail.
Pricing signal: Free trial, then subscription. Verify live before you commit.
Best for: Writers, docs, and engineers who think faster than they type.
Skip if: You share a noisy open office.
- Lane: Productivity
- Competes with: Apple Dictation, Whisper-based tools
- PH rank: Consistent top-10 daily in June 2026
Security
10. Astra Autonomous Pentest
Verdict: Watch
Astra is an AI agent that finds, validates, and fixes security vulnerabilities. It is a narrow but important vertical: offensive security. The agent runs scans, writes proof-of-concept exploits, and suggests patches.
Pricing signal: Enterprise-focused. No transparent self-serve pricing yet. Contact sales.
Best for: Security teams that want to augment human pentesters, not replace them.
Skip if: You need a compliance report from a human-certified firm for your auditor.
- Lane: Security / Agents
- Competes with: Traditional pentest firms, Burp Suite extensions
- PH rank: Top 3 daily launch June 4, 2026
11. Lumo Studios
Verdict: Watch
Lumo Studios is a new AI deck builder that launched on Product Hunt in early June 2026. It generates slides with a focus on narrative flow and speaker notes, not just layouts. Too new for a full verdict, but worth watching because the slides lane is crowded and any new entrant with a different angle matters.
Pricing signal: Free tier with limited decks. Paid tier for export and custom branding. Verify live.
Best for: Speakers who care about story arc, not just pixel-perfect slides.
Skip if: You need PPTX export today. Test before you commit.
- Lane: Creators / slides
- Competes with: Gamma, Dokie, SlideAI, Pitch
- PH rank: Top 15 daily launch June 5, 2026
Design
12. Ideogram 4.0
Verdict: Use
Ideogram 4.0 shipped in early June 2026 with open-weight design-ready image generation and layout control. The pitch is simple: generate images that look like finished design assets, not raw prompts. Layout control means you can specify regions and text placement.
Pricing signal: Free tier with daily credits. Paid tiers for commercial use and higher resolution. Verify live on ideogram.ai/pricing.
Best for: Designers and marketers who need social graphics, ads, and thumbnails without opening Figma.
Skip if: You need photorealistic human portraits. Ideogram excels at graphic design, not realism.
- Lane: Creators / design
- Competes with: Midjourney, DALL-E, Recraft
- PH rank: Top 5 daily launch June 5, 2026
13. OpenCode
Verdict: Use (builders, open-source agent)
OpenCode is an open-source AI coding agent that runs locally or in the cloud. It handles multi-file edits, terminal commands, and test execution through a clean desktop UI or CLI. Because it is open source, you can self-host it, audit the agent logic, and wire your own API keys for any model provider.
Pricing signal: Free and open source (Apache 2.0). You bring your own API keys (OpenAI, Anthropic, OpenRouter, or local models via Ollama). No subscription wall. Cloud-hosted options exist if you do not want to run the server locally.
What we tested: One TypeScript refactor adding error handling to a REST API. OpenCode read the repo, proposed changes across three files, ran tests, and surfaced one breaking change in the diff viewer. We accepted the safe hunks and rejected the breaking one.
Best for: Developers who want an open-source agent they can audit, self-host, and pair with cheap model lanes like DeepSeek V4 or local Qwen.
Skip if: You want a polished commercial IDE with one-click deploys and managed billing. OpenCode is a power-user tool.
- Lane: Builders
- Competes with: Cursor, Devin Desktop, Claude Code, Codex Desktop
- Repo: github.com/opencode (verify live; name collisions exist)
Skipped this week (logged, not forgotten)
| Tool | Reason |
|---|---|
| Statim / 13F.chat | Vertical finance; off-lane |
| PettiChat | Gadget |
| Lalals | Music niche |
| Clearfy | Humanizer policy skip |
| Manus / SlideAI / VmakeAI | Covered in June Week 1 radar; no duplicate slots |
| Genspark | Week 2; full review pending |
| SocialEcho 2.0 | Social media niche; radar only if queries rise |
| Databox MCP | MCP tooling; covered when MCP cluster ships |
| Vokal | Team collab; overlaps Slack/Notion lane |
| Minimi | Ambient memory; too new for verdict |
| Perplexity Pages | Covered in earlier radars; not a repeatable deliverable tool |
| ChatGPT Agent mode | Bundled into Plus; not a standalone product worth a radar slot |
Next week watch list
- June Week 4 roundup if search demand stays strong for weekly formats
- SlideAI credit promo changes
- Roger Google Ads agent if queries rise
- Lumo Studios deep test if PPTX export lands
- Mina CRM integration review if enterprise tier opens
- OpenCode deep test if local model hosting expands
Lane balance (Week 3)
| Lane | Tools |
|---|---|
| Slides | Dokie, Gamma, Kimi, Lumo Studios |
| Builders | Devin Desktop, Cursor, Codex Desktop, GitHub Copilot Workspace, OpenCode |
| Agents | Mina, Astra |
| Productivity | Wispr Flow |
| Design | Ideogram 4.0 |
Week 3 is for readers tracking builder lane tools and early June launches. We rescanned Product Hunt, AIxploria, and TAAFT on June 5, 2026. Start with the at-a-glance table, then open SlideAI review if you want SlideAI’s credit table side by side with Dokie and Gamma.
Free tier worksheet
| Tool | Free gives you | Paid unlock we saw | Hidden cost |
|---|---|---|---|
| Dokie | Generation on free credits | Plus for PPTX export | ~100 credits/deck |
| Gamma | Monthly AI credits | PPTX export | Credits burn on regen |
| Devin Desktop | Limited agent runs | Premium models | Premium model surcharge |
| Cursor | Hobby agent cap | Pro agent pool | Background agent quota |
| Kimi | Chat usage | PPTX unclear | Locale-specific UI |
| Codex Desktop | Daily agent limits | More runs + larger context | API usage may bill extra |
| Copilot Workspace | Public repos only on free | Business/Enterprise tiers | Seat minimums on paid |
| OpenCode | Free, open source | Cloud hosting optional | You bring your own API keys |
| Mina | Limited meeting minutes | CRM integrations | Bot compliance blocks |
| Wispr Flow | Trial period | Full dictation speed | macOS only |
| Astra | None visible | Enterprise contact sales | No self-serve |
| Lumo Studios | Limited decks | Export + branding | Too new to verify |
| Ideogram 4.0 | Daily credits | Commercial use + HD | Credit burn on retries |
Screenshot plan pages the day you pitch a client. Early June numbers can change by July.
Who should read this post first
| Reader | Start here |
|---|---|
| Agency PM | Dokie + Gamma + Ideogram rows |
| Engineering lead | Devin Desktop + Cursor + Codex Desktop + OpenCode + Astra rows |
| GitHub-native team lead | Copilot Workspace + Cursor rows |
| Sales rep | Mina + Wispr Flow rows |
| Finance ops | Skip to Excel guide on calendar |
Bottom line: Watch Dokie until export and credit math work on your templates (see full review). Use Gamma (link-first), Devin Desktop, Cursor, Codex Desktop, GitHub Copilot Workspace, OpenCode, Mina, Wispr Flow, and Ideogram 4.0 after you verify plan pages. Stay on Watch for Kimi PPT, Astra, and Lumo Studios until export, billing, and reliability match your client SLA.
Frequently asked
5 questionsWhat happened to Week 1 and Week 2?
Week 1 covered Manus, Dokie, SlideAI, and Claude Opus 4.8. Week 2 covered Genspark, OpenClaw, Roger, and Lovable. Week 3 focuses on builder lane tools that launched or rebranded in early June 2026.
Is Dokie AI in this week's list?
Dokie gets a full review slot on the editorial calendar. We include it here with export notes and link to the dedicated review when published.
Did SlideAI pricing change in June 2026?
We rechecked slideai.net on June 5, 2026. Free daily credits still apply. Paid packs still start at $1.99 for 20 credits per our review. Verify live before client work.
What is Devin Desktop doing on the radar?
Devin Desktop (formerly Windsurf) rebranded on June 2, 2026. It is a builder-lane IDE with rising vs Cursor queries. The full comparison post is already live.
Where is Claude Opus 4.8 this week?
On the models hub only unless Anthropic ships a new SKU. No repeat slot needed when release notes are unchanged.
More in Radar
View all
New AI Tools 2026 (June Week 1): 7 Tested, Ranked and Reviewed
New AI tools 2026 for June Week 1: 7 launches tested with pricing, Use Watch Skip verdicts, and links to full reviews. Updated weekly.
Radar

New AI Tools 2026 (June Week 2): 7 New Launches Reviewed
New AI tools 2026 (June week 2): seven launches reviewed with pricing, free tiers, and Use Watch Skip verdicts. Updated weekly.
Radar

AI Tools We Stopped Using in 2026 (7 Failed Experiments)
AI tools we stopped using in 2026: seven failed workflows, why we churned, and what we use instead. Honest anti-hype report.
Radar
More stories
View all
GLM-5.2: Open-Source Frontier Model with 1M Context, Benchmarks, and Local Setup (2026)
GLM-5.2 from Zhipu AI is a 744B open-weight model under MIT license. Benchmarks, pricing, local setup with vLLM and llama.cpp, and how it compares to Claude Opus 4.8 and GPT-5.5.
Models

Kimi K2.7 Code (2026): 1T MoE Coding Model, Benchmarks & Pricing
Kimi K2.7 Code: 1T open-source coding model from Moonshot AI, 32B active MoE, preserve_thinking mode, benchmarks vs GPT-5.5 and Claude Opus.
Models

MiniMax M3 Open Source (2026): 428B Model, 1M Context & Benchmarks
MiniMax M3: 428B open-weights model, 1M context via sparse attention, native multimodal input, competitive coding benchmarks, and 10x cheaper than GPT-5.5.
Models

US Government Blocks Anthropic Fable 5 & Mythos 5 (2026)
US government ban on Anthropic: Commerce Dept ordered suspension of Fable 5 & Mythos 5 on June 12, 2026. Full timeline of the 4-month feud.
Models