{"data":{"items":[{"id":"e168ad36-d9a0-4833-9026-7d86f2e777fb","excerpt":"PSA: Anthropic clarified the OpenClaw ban. You can still use your $20/mo sub (and avoid $1,500+ API bills). Here is the official CLI workaround — A few weeks ago, Anthropic announced they were banning third-party tools like OpenClaw from using Claude Pro or Max subscriptions, forcing everyone over to the pay-per-token ","url":"https://www.reddit.com/r/openclaw/comments/1svmq20/psa_anthropic_clarified_the_openclaw_ban_you_can/","role":"pain","weight":1.2441236,"occurredAt":"2026-04-25T20:17:50.000Z","sourceKey":"reddit","sourceName":"Reddit","credibility":0.62,"venue":"openclaw","intent":"problem_report","painScore":0.5083333,"sentiment":-0.33333334,"confidence":0.82483333,"matchedPatterns":["workaround","manual_process","product:anthropic"],"statement":"Here is the official CLI workaround.","title":"PSA: Anthropic clarified the OpenClaw ban. You can still use your $20/mo sub (and avoid $1,500+ API bills). Here is the official CLI workaround","body":"A few weeks ago, Anthropic announced they were banning third-party tools like OpenClaw from using Claude Pro or Max subscriptions, forcing everyone over to the pay-per-token API.  \n If you run heavy agentic workflows like I do, you probably watched your monthly bill skyrocket, I was on track to hit over $1,500 a month just running Opus 4.6.  \nBut Anthropic staff just clarified their policy, and the community completely missed the nuance.  \n**You can still legally run OpenClaw on your existing subscription.**   \nHere is the breakdown of the new rules and how to configure your stack today.\n\n**What is actually banned:** Tools that extract your OAuth token from your local Claude Code install and call the Anthropic API directly, pretending to be Claude Code. Anthropic hates this because it skips their prompt caching and puts a massive load on their servers without them getting the infrastructure benefits.  \n**What is officially sanctioned:** Shelling out to the local Claude CLI binary. Instead of extracting tokens, OpenClaw literally invokes the `claude` command on your machine with the `-p` flag. Because Claude Code is running the request natively, it handles all the authentication and caching exactly as Anthropic intended. OpenClaw team confirmed with Anthropic staff that this is an official green light.\n\n**How to configure it (The 1-Minute Fix):** You can either have Claude Code walk you through switching back to the OAuth token, or you can manually edit your OpenClaw JSON file.\n\n1. Open your OpenClaw JSON config.\n2. Find the provider block.\n3. Set your provider to `claude-cli`.\n4. Set the mode to `oauth`. The next time you start a session, it will fire off the `claude` command on your Mac and route through your subscription rather than charging you per-token.\n\n**⚠️ The Catch (Read before you scale):** Your Pro/Max plan is not unlimited. It has a finite usage cap (usually 5 hours), and if you route all your automated background cron jobs through it, you will hit that cap in 10 minutes and your agent will die.\n\n**Pro-Tip Architecture:** Save your Claude CLI subscription strictly for your main planning, strategy, and chat sessions that actually need Opus 4.7 quality. Take all of your automated background tasks and cron jobs and route them to cheaper alternatives, like OpenAI's GPT-5.4 API, or the dedicated 10−28/month coding CLI plans from GLM, Qwen, or MiniMax.\n\nLet me know if you run into any JSON errors setting this up!","offTopic":false},{"id":"e8eb902f-5548-43c0-91c9-c98f822a0e0a","excerpt":"Anthropic killed your Claude subscription for OpenClaw. Here's what to do now. — So if your OpenClaw just... stopped working around April 4 and you've been losing your mind trying to figure out why -- it's not you. Anthropic pulled the plug.\n\n**What happened:**\n\nAnthropic blocked all Claude Pro and Max subscriptions fr","url":"https://www.reddit.com/r/OpenClawUseCases/comments/1sho0sz/anthropic_killed_your_claude_subscription_for/","role":"demand","weight":1.0767456,"occurredAt":"2026-04-10T14:24:13.000Z","sourceKey":"reddit","sourceName":"Reddit","credibility":0.62,"venue":"OpenClawUseCases","intent":"alternative_search","painScore":0.35341814,"sentiment":-0.2,"confidence":0.79557496,"matchedPatterns":["switching_from","free_tier","product:anthropic"],"statement":"**Consider a refund.** If you're switching away from Claude entirely, Anthropic is offering full subscription refunds since the terms changed.","title":"Anthropic killed your Claude subscription for OpenClaw. Here's what to do now.","body":"So if your OpenClaw just... stopped working around April 4 and you've been losing your mind trying to figure out why -- it's not you. Anthropic pulled the plug.\n\n**What happened:**\n\nAnthropic blocked all Claude Pro and Max subscriptions from powering OpenClaw and basically every other third-party agent tool. April 4, 12pm PT. If you were running Claude through OAuth or setup-token... yea that's dead now.\n\nThis wasn't a bug or some accident. Boris Cherny (Head of Claude Code at Anthropic) said it straight up -- subscriptions weren't built for the kind of usage these tools generate. And I mean... he's not wrong? One Register writer was getting $236 worth of tokens from a $20 subscription. Some Max users were burning through thousands of dollars worth of compute *daily* on a $200/month plan. Anthropic looked at that and said nah.\n\nPeter Steinberger (OpenClaw creator, now at OpenAI) tried to negotiate with them. Him and Dave Morin. Best they got was a one week delay. Steinberger accused Anthropic of copying popular OpenClaw features into Claude Code and then locking out the open source alternative. I mean... the timing is suspicious but idk, make of that what you will.\n\n**What still works:**\n\nYour Claude subscription still works for Claude.ai, Claude Code, and Cowork. Just not for OpenClaw or any third-party agent.\n\n**Your options:**\n\n**Option 1: Direct Anthropic API key** \\-- This is probably what most people should do. Go to console.anthropic.com. Generate an API key. Load credits.\n\nbash\n\n    openclaw onboard\n    # choose: Anthropic API key\n    # paste your key\n\nOr just set it in your config:\n\njson\n\n    {\n      \"env\": {\n        \"ANTHROPIC_API_KEY\": \"sk-ant-...\"\n      }\n    }\n\nYou pay per token now. No surprises. Sonnet 4.6 at moderate personal use runs maybe $20-40/month. Heavy business use more like $60-120. Not cheap but at least it's predictable and you can see exactly what you're spending in the Anthropic console.\n\nThere's also this \"Extra Usage\" thing where you keep your subscription and OpenClaw usage gets billed separately on top. Go to your Claude account settings and enable Extra Usage. Your agent should reconnect. Honestly though... this is confusing and potentially more expensive than just going straight API. It exists but I'd stick with the API key.\n\n**Option 2: Ditch Claude for daily use**\n\nThis is what a LOT of people are doing right now and honestly it's not a bad move:\n\n* OpenAI Codex OAuth still works (for now). GPT-5.4 through your $23 Plus plan. But dude... if Anthropic did it, OpenAI might follow. Don't get too comfortable.\n* Minimax M2.7 at $10-20/month. Multiple people in the community calling it \"90% of Sonnet at 10% of the cost.\"\n* GLM 5.1 getting a lot of love. \"Closest to Claude\" according to a bunch of people I've talked to.\n* Ollama local for truly $0. Qwen3.5 or GLM-4.7 on your own hardware.\n* OpenRouter free tier. 30+ free models including Nemotron Ultra and Llama 3.3 70B.\n\n**Option 3: Hybrid (what I'd actually do)**\n\nDaily default: Minimax or GLM 5.1. Cheap, handles 70% of tasks fine.\n\nFallback: Sonnet 4.6 on API for the complex stuff.\n\nFloor: Ollama local so your agent doesn't just die at 2am because some provider is down.\n\nMonthly cost: $20-50 depending on Sonnet usage. Compared to the \"free\" ride everyone was getting before, yes it costs money now. Compared to the $200 Max plan... way cheaper and more predictable.\n\n**Before you do anything else tho:**\n\n**Claim your free credit.** Anthropic is giving everyone a one-time credit equal to your monthly plan price. Redeem it in your Claude account settings. Expires April 17. That's like... next week. Do not sleep on this.\n\n**Grab the 30% bundle discount.** If you're staying with Claude, you can pre-purchase Extra Usage bundles at up to 30% off right now.\n\n**Consider a refund.** If you're switching away from Claude entirely, Anthropic is offering full subscription refunds since the terms changed. Might as well.\n\nLook... community estimates from the Reddit threads suggest roughly 60% of active OpenClaw sessions were running on subscription credits. That's a massive chunk of the user base that just got hit with real costs for the first time.\n\nThe $20/month all-you-can-eat era for AI agents is over. Not just Anthropic. Google did the same thing with Gemini CLI back in February. OpenAI will probably follow eventually. The subscription arbitrage model is closing across the board.\n\nThe people who set up model routing with cheap defaults and paid fallbacks months ago? Nothing changed for them today. The people who built everything around one provider's subscription? Scrambling.\n\nBuild your setup so no single provider's pricing decision can break you overnight.","offTopic":false},{"id":"9c71b701-2248-4290-a526-bccdc257ce4b","excerpt":"Built an OpenClaw alternative that wraps Claude Code CLI directly & works with your Max subscription — Hey everyone. I've been running OpenClaw for about a month now and my API costs have been creeping up to the point where I'm questioning the whole setup. Started at \\~$80/mo, now consistently $400+ with the same workl","url":"https://www.reddit.com/r/openclaw/comments/1runkql/built_an_openclaw_alternative_that_wraps_claude/","role":"request","weight":0.9896267,"occurredAt":"2026-03-15T19:42:01.000Z","sourceKey":"reddit","sourceName":"Reddit","credibility":0.62,"venue":"openclaw","intent":"feature_request","painScore":0.36,"sentiment":0.14285715,"confidence":0.7276667,"matchedPatterns":["missing_feature","product:anthropic"],"statement":"The gateway just adds what they're missing: routing, cron scheduling, messaging integration, and a multi-agent org system.","title":"Built an OpenClaw alternative that wraps Claude Code CLI directly & works with your Max subscription","body":"Hey everyone. I've been running OpenClaw for about a month now and my API costs have been creeping up to the point where I'm questioning the whole setup. Started at \\~$80/mo, now consistently $400+ with the same workload ( I use Claude API as the main agent ).\n\nSo I built something different. Instead of reimplementing tool calling and context management from scratch, I wrapped Claude Code CLI and Codex behind a lightweight gateway daemon. The AI engines handle all the hard stuff natively including tool use, file editing, memory, multi-step reasoning. The gateway just adds what they're missing: routing, cron scheduling, messaging integration, and a multi-agent org system.\n\nThe biggest win: because it uses Claude Code CLI under the hood, it works with the $200/mo Max subscription. Flat rate, no per-token billing. Anthropic banned third-party tools from using Max OAuth tokens back in January, but since this delegates to the official CLI, it's fully supported.\n\n**What it does:**  \n• Dual engine support (Claude Code + Codex)  \n• AI org system - departments, ranks, managers, employees, task boards  \n• Cron scheduling with hot-reload  \n• Slack connector with thread-aware routing  \n• Web dashboard - chat, org map, kanban, cost tracking  \n• Skills system - markdown playbooks that engines follow natively  \n• Self-modification - agents can edit their own config at runtime\n\nIt's called **Jinn**: [https://github.com/hristo2612/jinn](https://github.com/hristo2612/jinn)","offTopic":false},{"id":"8c418dbb-9984-4f48-9990-a5111a0c15af","excerpt":"Anthropic just kicked OpenClaw users off their subscription. Here’s what to do — \n\nIf you’re reading this, you probably got the email from Anthropic. Starting today, your Claude Max subscription no longer covers OpenClaw or any other agents. Every agent message is now billed separately as “Extra Usage” on top of your s","url":"https://www.reddit.com/r/AskClaw/comments/1scxbps/anthropic_just_kicked_openclaw_users_off_their/","role":"pricing","weight":0.9503116,"occurredAt":"2026-04-05T07:26:07.000Z","sourceKey":"reddit","sourceName":"Reddit","credibility":0.62,"venue":"AskClaw","intent":"pricing_complaint","painScore":0.36,"sentiment":1,"confidence":0.69875854,"matchedPatterns":["paying_monthly","product:anthropic"],"statement":"If you’re paying $200/month and running OpenClaw agents, you’re now paying that subscription plus raw API costs for every token your agents consume.","title":"Anthropic just kicked OpenClaw users off their subscription. Here’s what to do","body":"\n\nIf you’re reading this, you probably got the email from Anthropic. Starting today, your Claude Max subscription no longer covers OpenClaw or any other agents. Every agent message is now billed separately as “Extra Usage” on top of your subscription.\n\n# What changed\n\nAnthropic has narrowed the scope of its Claude subscriptions. They now only apply to first-party surfaces: [Claude.ai](http://claude.ai/), Claude Code, and the mobile app. Everything else is treated as a “third-party harness” and billed separately on a per-token basis.\n\nIn practice, this means your Max plan no longer covers agent usage. If you’re paying $200/month and running OpenClaw agents, you’re now paying that subscription plus raw API costs for every token your agents consume.\n\n# Your options\n\n# 1. Stay on Max and pay extra\n\nYou’re probably looking at $300-500/month total depending on how many agents you run. Makes sense if you also use [Claude.ai](http://claude.ai/) directly a lot. Otherwise the math doesn’t add up.\n\n# 2. Switch to the Anthropic API\n\nDrop the subscription and run your agents on a straight API key from [console.anthropic.com](http://console.anthropic.com/). Here is the cost of some Claude models:\n\nhttps://preview.redd.it/x0rbbdh5qbtg1.png?width=824&format=png&auto=webp&s=c4a29a04161fe693fa05cb31d1da68413fe82660\n\n\n\n* Opus 4.6: $5 input / $25 output\n* Sonnet 4.6: $3 input / $15 output\n* Haiku 3: $0.25 input / $1.25 output\n\nMost people running a couple of agents will spend $200-400/month on API tokens, roughly the price of a Max plan.\n\n# 3. Switch providers\n\nOpenClaw isn’t tied to Claude. You can run it with OpenAI, Gemini, Kimi, or open-source models via Ollama.\n\nClaude is still strong, but Anthropic is tightening access. Others are more agent-friendly, especially OpenAI and players like Minimax already building for integrations.\n\n# 4. Use intelligent routing\n\nInstead of picking one model and sending everything to it, [Manifest](https://manifest.build/) routes each request to the right model automatically. Simple prompt goes to Haiku. Complex tasks goes to Sonnet. Each task gets the cheapest model that can handle it.\n\nIn practice, if you’re spending $200–400/month using API key, routing typically brings that down to around $60–120. Same idea on Max plan: instead of stacking usage on top of your subscription, you keep it under control.\n\n# What to do right now\n\nAnthropic is offering free API credits to eligible subscription users. Between $20 and $200 depending on your account. You have until April 17th to claim them.\n\nhttps://reddit.com/link/1scxbps/video/zkzle9mjpbtg1/player\n\nHere’s how to grab them and start routing:\n\n1. Go to [app.manifest.build](https://app.manifest.build/) and set up your agent if you haven’t already\n2. Click on the Routers tab, then “Add a router”\n3. Inside the router, click “Add a provider” and from the Subscription tab, select Anthropic\n4. You’ll see a button to claim your credits from Claude. Click it. It takes you to an Anthropic page where you can claim them. If you get a 404, your account isn’t eligible\n\nhttps://preview.redd.it/wxjkxplopbtg1.png?width=1026&format=png&auto=webp&s=ed19e7a0c9474c467f257b63f8898225c938b226\n\n\n\n1. Come back to Manifest with the modal still open\n2. Connect your Anthropic subscription by running `claude setup-token` in your terminal\n3. It opens a Claude tab in your browser. Authorize it and close the tab.\n4. A token appears in your terminal. Copy it, paste it into Manifest in the “Setup Token” field, click Connect\n5. Manifest pre-selects the best models for each tier. You can adjust them if you want\n\n\n\nhttps://preview.redd.it/swver4dqpbtg1.png?width=2540&format=png&auto=webp&s=d3c3c974c27292500ac88b463fdd26c43d86fde4\n\nThat’s it. Your credits last way longer because you’re not burning Opus tokens on everything.\n\n# The bigger picture\n\nPick the right plan for how you actually work. If you use [Claude.ai](http://claude.ai/) a lot, Max still makes sense. If you mostly run agents, a straight API key will cost you less.\n\nAnd pick the right model for each task. Routing fixes this automatically and smartly. As agent usage grows, it stops being optional.\n\nTo finish, if you’re on a Claude subscription, claim your free credits before April 17th. Between $20 and $200 sitting there for free.","offTopic":false},{"id":"3f8532c2-754e-42c7-ab34-4d3e1e78e3fd","excerpt":"[Showcase] Anthropic just kicked OpenClaw users off their subscription. Here’s what to do — If you’re reading this, you probably got the email from Anthropic. Starting today, your Claude Max subscription no longer covers OpenClaw or any other agents. Every agent message is now billed separately as “Extra Usage” on top ","url":"https://www.reddit.com/r/clawdbot/comments/1sddx0q/showcase_anthropic_just_kicked_openclaw_users_off/","role":"pricing","weight":0.9124171,"occurredAt":"2026-04-05T20:19:02.000Z","sourceKey":"reddit","sourceName":"Reddit","credibility":0.62,"venue":"clawdbot","intent":"pricing_complaint","painScore":0.36,"sentiment":1,"confidence":0.6708949,"matchedPatterns":["paying_monthly","product:anthropic"],"statement":"If you’re paying $200/month and running OpenClaw agents, you’re now paying that subscription plus raw API costs for every token your agents consume.","title":"[Showcase] Anthropic just kicked OpenClaw users off their subscription. Here’s what to do","body":"If you’re reading this, you probably got the email from Anthropic. Starting today, your Claude Max subscription no longer covers OpenClaw or any other agents. Every agent message is now billed separately as “Extra Usage” on top of your subscription.\n\n# What changed\n\nAnthropic has narrowed the scope of its Claude subscriptions. They now only apply to first-party surfaces: Claude.ai, Claude Code, and the mobile app. Everything else is treated as a “third-party harness” and billed separately on a per-token basis.\n\nIn practice, this means your Max plan no longer covers agent usage. If you’re paying $200/month and running OpenClaw agents, you’re now paying that subscription plus raw API costs for every token your agents consume.\n\n# Your options\n\n# 1. Stay on Max and pay extra\n\nYou’re probably looking at $300-500/month total depending on how many agents you run. Makes sense if you also use Claude AI directly a lot. Otherwise the math doesn’t add up.\n\n# 2. Switch to the Anthropic API\n\nDrop the subscription and run your agents on a straight API key from [console.anthropic.com](https://console.anthropic.com/). Here is the cost of some Claude models:\n\nhttps://preview.redd.it/x0rbbdh5qbtg1.png?width=824&format=png&auto=webp&s=c4a29a04161fe693fa05cb31d1da68413fe82660\n\n* Opus 4.6: $5 input / $25 output\n* Sonnet 4.6: $3 input / $15 output\n* Haiku 3: $0.25 input / $1.25 output\n\nMost people running a couple of agents will spend $200-400/month on API tokens, roughly the price of a Max plan.\n\n# 3. Switch providers\n\nOpenClaw isn’t tied to Claude. You can run it with OpenAI, Gemini, Kimi, or open-source models via Ollama.\n\nClaude is still strong, but Anthropic is tightening access. Others are more agent-friendly, especially OpenAI and players like Minimax already building for integrations.\n\n# 4. Use intelligent routing\n\nInstead of picking one model and sending everything to it, [Manifest](https://manifest.build/) routes each request to the right model automatically. Simple prompt goes to Haiku. Complex tasks goes to Sonnet. Each task gets the cheapest model that can handle it.\n\nIn practice, if you’re spending $200–400/month using API key, routing typically brings that down to around $60–120. Same idea on Max plan: instead of stacking usage on top of your subscription, you keep it under control.\n\n# What to do right now\n\nAnthropic is offering free API credits to eligible subscription users. Between $20 and $200 depending on your account. You have until April 17th to claim them.\n\nHere’s how to grab them and start routing:\n\n1. Go to [app.manifest.build](https://app.manifest.build/) and set up your agent if you haven’t already\n2. Click on the Routers tab, then “Add a router”\n3. Inside the router, click “Add a provider” and from the Subscription tab, select Anthropic\n4. You’ll see a button to claim your credits from Claude. Click it. It takes you to an Anthropic page where you can claim them. If you get a 404, your account isn’t eligible\n\nhttps://preview.redd.it/wxjkxplopbtg1.png?width=1026&format=png&auto=webp&s=ed19e7a0c9474c467f257b63f8898225c938b226\n\n1. Come back to Manifest with the modal still open\n2. Connect your Anthropic subscription by running `claude setup-token` in your terminal\n3. It opens a Claude tab in your browser. Authorize it and close the tab.\n4. A token appears in your terminal. Copy it, paste it into Manifest in the “Setup Token” field, click Connect\n5. Manifest pre-selects the best models for each tier. You can adjust them if you want\n\nhttps://preview.redd.it/swver4dqpbtg1.png?width=2540&format=png&auto=webp&s=d3c3c974c27292500ac88b463fdd26c43d86fde4\n\nThat’s it. Your credits last way longer because you’re not burning Opus tokens on everything.\n\n# The bigger picture\n\nPick the right plan for how you actually work. If you use Claude AI a lot, Max still makes sense. If you mostly run agents, a straight API key will cost you less.\n\nAnd pick the right model for each task. Routing fixes this automatically and smartly. As agent usage grows, it stops being optional.\n\nTo finish, if you’re on a Claude subscription, claim your free credits before April 17th. Between $20 and $200 sitting there for free.","offTopic":false},{"id":"271f8614-a46b-4289-82ec-5b2391d4ec7b","excerpt":"ArmyClaw = Make your Claude Code subscription 100x more productive. — # ArmyClaw: 24/7 Agents on Your Existing Claude Code Subscription\n\nWant 24/7 OpenClaw-style agents but on your existing Claude Code subscription? Meet **ArmyClaw**. Make your Claude Code subscription 100x more productive.\n\n## Why ArmyClaw Exists\n\nAnt","url":"https://www.reddit.com/r/AI_Agents/comments/1t29ign/armyclaw_make_your_claude_code_subscription_100x/","role":"pain","weight":0.8253204,"occurredAt":"2026-05-03T02:51:19.000Z","sourceKey":"reddit","sourceName":"Reddit","credibility":0.62,"venue":"AI_Agents","intent":"problem_report","painScore":0.42666668,"sentiment":-0.16666667,"confidence":0.5784956,"matchedPatterns":["manual_process","product:anthropic"],"statement":"No copy-pasting context between sessions.","title":"ArmyClaw = Make your Claude Code subscription 100x more productive.","body":"# ArmyClaw: 24/7 Agents on Your Existing Claude Code Subscription\n\nWant 24/7 OpenClaw-style agents but on your existing Claude Code subscription? Meet **ArmyClaw**. Make your Claude Code subscription 100x more productive.\n\n## Why ArmyClaw Exists\n\nAnthropic just blocked OpenClaw from piggybacking on your plan — they were extracting OAuth tokens and spoofing headers. Now if you want OpenClaw with Claude, you need API keys. Real API pricing. Thousands of dollars a month for what your flat-rate plan already covers.\n\n## How ArmyClaw Is Different\n\nArmyClaw takes a completely different approach:\n\n- Spawns the actual `claude` CLI binary as a subprocess\n- Authenticates through your legitimate claude login session\n- Orchestrates around the official tool\n- **No token theft. No header spoofing. No policy violation.**\n\nYour existing Pro or Max subscription powers everything — no API keys, no credits burned, no surprise bills.\n\n\n## Key Features\n\n### 🧠 Agents That Actually Talk to Each Other\nCross-chat collaboration with shared long-term memory. What one agent learns, every other agent can access. No copy-pasting context between sessions.\n\n### 💬 Group Brainstorming Rooms\n2–5 agents debate your problem Slack-style, not just respond to you.\n\n### 📱 Multi-Platform Control\nDrive any agent from Telegram, your browser, or the built-in terminal. Start a task on your laptop, finish it from your phone.\n\n### 🎭 Unlimited Personas\nPer role, project, or client. Color-coded, filterable, each with their own personality and expertise.\n\n### 🔱 Conversation Forking\nFork any conversation with the last 200 turns inherited. The new agent already knows what the parent knew.\n\n### ⏰ Scheduled Routines Per Agent\nMorning PR reviews, hourly monitoring, nightly reports. Survives restarts.\n\n### 🔄 Crash Recovery\nDetects interrupted sessions and self-resumes with a synthetic wake-up. You see no hiccup.\n\n### 📸 Workspace Snapshots\nTime-travel your entire workspace. Roll back before risky experiments.\n\n### 🔌 Swap to Any Model Provider\nOpenRouter, DeepSeek, Kimi, GLM, Ollama, fully offline. Two env vars, done.\n\n### 🛠️ Built-In Tools\nTerminal, file explorer, artifact canvas, voice input, full-text search across all agents, 10 themes.\n\n---\n\nWould love feedback, issues, and PRs.","offTopic":true}],"breakdown":[{"sourceKey":"reddit","sourceName":"Reddit","count":6}],"total":6}}