BlockRunAI

BlockRun.AI - Block = Blockchain/ on Blockchain - Run = Execute/Operate - AI = AI Agent

The @base Ecosystem Tier List - Week 28 Base gives builders a direct path from product to users, liquidity, and revenue. The builders defining this week across the ecosystem: OG - @AerodromeFi - @Morpho - @AskVenice - @virtuals_io - @bankrbot - @Uptopia_xyz - @o1_exchange - @ribbita2012 - @Toshi - @degentokenbase Based - @avantisfi - @POVMarket - @opal_dex - @clanker_world - @BlockRunAI - @gitlawb - @AskSurplus - @OpenGradient - @reppo - @shadowcleague - @aeonframework - @HydrexFi Rising - @officialbunnyos - @StrikeRobot_ai - @dphnAI - @lienfiapp - @dominationfi - @Wach_AI - @AlienBaseDEX - @PixieChess - @ClawBankHQ - @miroshark_ - @PRXVTai - @sibylcap - @Litebeam_xyz - @Trueo_ Early - @rwagmicom - @playhunchxyz - @vibestarterxyz - @1clawAI - @helixaxyz - @ForecastFDN - @liminalcash - @blocktronicsxyz - @PipelineMarkets - @nook_platform Which active builder did we miss this week?

Peter want us to merge builder code....

Try - openai/gpt-5.6-sol — $5 / $30 per 1M - openai/gpt-5.6-terra — $2.50 / $15 - openai/gpt-5.6-luna — $1 / $6 with BlockRun

Top 10 x402 Projects on @base 1. @AskVenice - private model inference through x402, giving agents paid access without API keys or billing accounts. 2. @bankrbot - LLM Gateway and agent tools where builders can access models, apps, and paid endpoints through agent-native payments. 3. @merit_systems - x402 analytics and discovery engine powering visibility across sellers, buyers, servers, and payment activity. 4. @heysealai - Hacks marketplace where users can build agent skills from prompts and earn when other agents use them. 5. @Litebeam_xyz - service routing for agents, helping them find the right vendor, validate inputs, and pay only for usable calls. 6. @OpenGradient - verifiable LLM inference with TEE-backed execution, proofs, and pay-per-request settlement. 7. @Wach_AI - verification and dispute rails for agent commerce, including merchant checks before x402 credit finalization. 8. @BlockRunAI - pay-per-call AI gateway for models, data, runtime, RPC, voice, image, video, and other agent services. 9. @Uptopia_xyz - GameAI loops with 180K+ x402 transactions across autonomous gameplay and onchain actions. 10. @useOttoAI - paid agent endpoints with 3,400+ calls from 148 unique buyers in one week. x402 is the payment rail for the agent economy on base.

6 months ago, @BlockRunAI was one question: what if AI agents could pay for exactly what they use — per request, in real time? The answer so far: 12M+ transactions in the last 30 days, settled in @USDC on @base. Grateful for the devs who shipped with us, the community that believed early, and the partners betting on agent-native payments. We set out to build the rails for the Inference Economy. Six months in, they're carrying real traffic. Onward.

@solana_devs thank @solana_devs for this opportunity

We'll be back with a lot more soon.

San Francisco is far from dead and still has one of the highest concentrations of developer talent and ambitious builders. The excitement from our events last week proved one thing: SF wants more Solana. Appreciate everyone who came to Demo Night, and the teams that presented: @unruggable_io @Vishwa_lab @flovia402 @clawpumptech @BlockRunAI

@a12321xyz @ZayyanuZay60173 And we will see SeeDance2.5 come soon

@ZayyanuZay60173 @BlockRunAI - Seedance 2.0 brings exciting advancements to AI cinematic content creation.

🚨 Cinematic AI just leveled up. Seedance 2.0 · 4K is LIVE on @BlockRunAI MCP! 🎥🔥 One prompt. Your agent handles the rest. Stunning cinematic 4K output with seamless USDC pay-per-clip. The era of AI-powered content creation is moving at full speed, and @BlockRunAI MCP is making it ridiculously easy. Prompt. Generate. Dominate. Try it here github.com/BlockRunAI/blo…

use @BlockRunAI for SeeDance 2.0, and we will have SeeDance 2.5 come out soon

sick

🎬 Seedance 2.0 · 4K is now live on BlockRun MCP！ Prompt in your agent, cinematic 4K out. Pay-per-clip in USDC. Try it 👇 github.com/BlockRunAI/blo…

Integrity matters

Big @BlockRunAI week across the stack. Model lineup got a strong refresh: Claude Sonnet 5 is live, GLM 5.2 landed, Kimi K2.7 is in the flow, and the catalog/picker side got cleaned up so users can find the newer models without guessing IDs. Codex support moved forward too: clawrouter-codex lets Codex CLI/Desktop run on BlockRun models, with direct x402 mode, blockrun/auto routing, web search, a local dashboard, wallet funding view, spend chart, and model picker. Try it out here : github.com/BlockRunAI/cla… Openclaw and Hermes users also got a smoother Telegram picker. GLM 5.2 and Sonnet 5 and few new models are now in the UI selector, ordering is cleaner, and the picker is staying closer to the current ClawRouter model lineup instead of showing stale or noisy entries. Tool calling got less cursed. We fixed Gemini tool-call leaking, GPT 5.4 plain-text tool-call leaking, and other cases where models returned tool calls in weird text formats instead of proper structured calls. Agents now execute tool calls more smoother and cleaner instead of dumping raw tool-call text into chat. ClawRouter also got reliability work: - fixed stale payment pre-auth reuse on larger requests - cleaned up /v1/models so delisted/redirect aliases don’t pollute clients - dependency/security cleanup brought audit state to 0 vulns - update/install cleanup around stale OpenClaw metadata Image generated with blockrun MCP

Incredibly excited to have @bc1beat from @BlockRunAI join us for our Web3 x AI Space tomorrow. Grateful she's joining us in light of the @hf0 accelerator and its accompanying demo days this week. Vicky is the former Head of Data Science and Analytics at @circle, and currently runs one of the leading Web3 x AI projects by adoption. BlockRun has processed more than 10.42M cumulative transactions across ~988 buyers as of July 1st, 2026. BlockRun provides a pay-per-call AI gateway across models, data, and runtime through one API endpoint, 60+ models accessible across chat, image, video, and audio inputs and outputs. The wider vision encompasses budget controls, usage analytics, and team management. BlockRun's business model is unique: ~5% margin at settlement, with no training on or retention of the data inputs and prompts provided to a model. Model orchestration, run-time inference, and the wider AI inference market continue to accelerate. Token optimization and model orchestration remain relevant in light of recent significant raises across @OpenRouter, @baseten, and @AskVenice. BlockRun operates in this market, and continues to lead x402-volume adoption. Join us for our conversation tomorrow at 2PM ET tomorrow, 11AM PT.

🎙️X Spaces: Web3 x AI 📅Thursday, 2nd July ⏰ 11 AM PT | 2 PM ET | 1 PM CST Moderator: @abdulalali Co-Host: @bc1beat of @BlockRunAI 🔗x.com/i/spaces/1mGPa… Mark your calendars and set a reminder – see you!

today I am asking my BlockRun-mcp to generate an image for me, anything github.com/BlockRunAI/blo… This is what it generated.

BlockRun have Sonnet 5 now blockrun.ai/models/anthrop…

Top AI / Robotics Plays on @base this week AI ■ @AskSurplus - spot inference marketplace saving users ~$15K/week, with nearly $800K/year in savings at current pace. ■ @dphnAI - Dolphin Network V2 is live for uncensored distributed inference, with 64.4B tokens generated in one day. ■ @Velvet_Capital - started training Velvet-1, an onchain intelligence model for wallets, flows, markets, and social signals. ■ @BlockRunAI - hit $100K MRR in x402 volume after Base Ecosystem Fund backing from @cbventures. ■ @gitlawb - ZERO ships this week, a Go-based coding harness built from scratch for faster AI coding workflows. ■ @Uptopia_xyz - Agentopia is coming with the first 100 agents helping builders test demand, users, and actions. ■ @miroshark_ - MiroShark Sim is live for pay-per-run market simulations from prompts, URLs, and documents. Robotics ■ @axisrobotics - Policy Checker is live, with browser-run robot policies, success heatmaps, and harder long-horizon tasks. ■ @opengotchi - AI desk pet running gotchiOS, programmable in one prompt, with v1 reservations live. ■ @shadowcleague - robot combat league building SCL around humanoid fights, creator moments, and live sports events.

Top 10 Projects of the Week on @base Big week for the Base ecosystem, especially around AI agents, RWA, DeFi, x402 payments, and onchain trading. 1. @virtuals_io Made Wall Street assets tradeable by AI agents, including tokenized stocks like SpaceX, Apple, Tesla, and NVIDIA 2. @avantisfi Reached a $50M OI cap on USD/JPY and activated RWA Phase 1A for zero-fee leveraged trading on FX pairs and metals 3. @BlockRunAI Secured investment from the Base Ecosystem Fund by @cbventures as one of the top AI agent payment platforms by USDC volume 4. @PancakeSwap Opened 15 high-yield Base liquidity pools with 80%+ APR and added deeper stock fundamentals for 500+ tokenized assets 5. Morpho Base highlighted Morpho’s lending scale, with billions in borrow and collateral volume already running onchain 6. @opensea Enabled AI agents to swap tokens, buy NFTs, and mint collections across chains through Base MCP 7. @Litebeam_xyz Enabled x402 payments across @bitrefill and 3,500+ AI service vendors through Base MCP 8. @Brickken Brought programmable asset tokenization infrastructure and x402 payments into Base MCP 9. @AskVenice Integrated private model inference with x402 support and launched Seedance 2.0 Mini for anonymous text-to-video generation 10. @anyone_market Launched a new prediction market platform where anyone can create markets and earn 50% of the fees via AI agents

July 1 Base daily recap What happened on @base over the last 24 hours. Base had a busy day across outage transparency, Coinbase AI efficiency, agent growth, x402 volume, onchain intelligence, robotics simulation and World Cup apps. NEWS @buildonbase published the June 25-26 outage post-mortem. The root cause has been fixed, and all funds remained safe. @brian_armstrong also shared how Coinbase cut AI spend nearly in half while token usage continued to grow. AI @virtuals_io crossed 50K agents launched. @dphnAI generated 64.4B tokens in one day, worth around $15.4K/day at OpenRouter floor rates, while @BlockRunAI crossed roughly $100K/month in x402 volume and is now averaging $4K-5K/day. @Velvet_Capital started training Velvet-1, its LLM for onchain intelligence. @Uptopia_xyz shared Agentopia plans for the first 100 agents to test traction, x402 swaps, feedback and security, while @Darks0l_ shipped ReMEM v0.19.0 with Codebase Graph memory. @aeonframework and @miroshark_ shipped 150 PRs last week, with MiroShark Sim now live and 174 paid x402 runs. @freecodexyz is building open-source repo shares for onchain trading, and @Root_Edge updated its Hyperliquid LLM trading UX with free inference and portfolio tools. ROBOTICS @axisrobotics expanded articulated-object simulation and long-horizon task generation. @opengotchi also got a breakdown on why AI minds may need physical bodies. ECOSYSTEM @sportfun published its WC2026 group stage recap, with Messi leading SDF points at 542. The bigger signal is how much of Base activity now sits around agents. Inference volume, x402 payments, agent launches, code memory, trading UX, simulations and robotics data are all moving at the same time. That makes Base feel less like a single app ecosystem and more like an execution layer for AI-native products. More agent flow. More onchain activity. Better infrastructure. Base keeps expanding.