Rightbrain AI

shiny new docs for the builders docs.rightbrain.ai/docs/gett…

We're back on X and we've been cooking

Opus 4.6 is now live on Rightbrain AI. Anthropic's most intelligent model. Ready to plug into production-ready agents and tools that run inside your existing platforms. Built for complex reasoning. Not just speed. Where nuance determines whether the output is useful or noise. Some inspiration: → Contract analysis that cross-references master agreements against amendments and side letters. Flags conflicts. Outputs structured risk assessments into your legal workflow. → Compliance reasoning that holds multiple regulatory frameworks in context at once (FCA, GDPR, internal policy) and returns pass/fail verdicts with specific citations. → Research synthesis that connects competitor data, market signals, and CRM context. Produces strategic briefs with non-obvious insights, not summaries. → Incident root cause analysis across logs, alerts, and metrics. Structured for PagerDuty or Jira. → Messy data classification that handles the edge cases lighter models get wrong. Confidence-scored. → Financial narratives from raw data that a CFO would actually use. Not templates. Deep analysis. What use case are you going to unlock?

🔬 AI Observability is key - built in audit logs give you a live stream of all your task runs, wherever they are deployed. Monitor AI tools that are integrated into workflows, apps or agents through one dashboard. - See failed runs and diagnose quickly - Compare tokens, cost and speed - Spot any anomalies across key metrics

Last week felt a little familiar 🎥 @GoogleDeepMind , @AnthropicAI and @OpenAI all dropped new models. Beyond the hype, our favourite take came from @simonw - 'I’ve mainly explored full GPT-5. My verdict: it’s just good at stuff.' GPT-5 may not represent a huge leap in comparison with other LLMs but seems to deliver the level of consistency for most biz use cases.

Here we go! We're excited to support the launch of OpenAI's open-weight model gpt-oss-120b, designed for powerful reasoning, agentic tasks, and versatile developer use cases. Now available on Rightbrain and ready to test across any text-based task, like code reviews, content moderation, and document summarisation. Best practices: 🔧 Define tools clearly — include parameters, output formats, and constraints 🧩 Break down tasks into step-by-step instructions to guide the model’s reasoning 📄 Use long context to enable richer workflows and cross-document analysis Head over to rightbrain.ai/ to try it for yourself.

⚡ Need for Speed 📷 Kimi K2 by Moonshot AI has been making waves with its combo of speed and performance, so we put it to the test against Gemini 2.5 Flash-Lite on an article analysis task.

🧠 Turn any prompt into a production-ready AI tool in 3 steps: 1️⃣ Find a repetitive task your team runs 2️⃣ Create your prompt with input variables (e.g., "Generate product listing from photo + {description}") 3️⃣ Use Rightbrain assistant for consistent results Then plug into your existing workflows via API or MCP 🚀

📽️A little teaser for our next major release - a single dashboard for managing AI tools. See requests, results, cost and latency stats for AI use across any app or agent. 💡 Why? We keep hearing from tech and biz leads that their teams are adopting AI but most of this happens without any real oversight. They want to enable their colleagues and customers but worry about the quality and consistency of outputs. ⏩ Build in Rightbrain, deploy via MCP wherever you need your tools to be.

⏩Gemini 2.5 Pro and 2.5 Flash are now live on Rightbrain, available for your new and existing production AI tools

🎁 Introducing Claude 3.7 Sonnet from the team at @AnthropicAI now available for testing on Rightbrain. 📸 Snapshot: > the first hybrid reasoning model - users can control how long the model thinks for, choosing from near-instant responses to extended reasoning > state-of-the art for coding - can complete tasks across the entire software development lifecycle > significant improvements in content generation, data analysis, and planning - able to understand nuance and tone to create more compelling content ⏩ We try it out on a simple task that generates an instant elevator pitch from our mates @descript's website. Equally applicable to any task that takes in unstructured data from a url as input and converts it to a pre-specified format.

📽️ Boost your LLM deployments with a consistent data structure. Here’s how it helps: • Model Flexibility: Swap models without changing your core logic. • Guaranteed Consistency: Every response meets your predefined structure. • Automated Validation: Catch invalid responses before production. • Frontend Reliability: Stable data means robust UI features. • Future-Proofing: Seamless upgrades & model swaps. Check out our quick demo to see it in action.

🔬 Diving Deeper into DeepSeek - rightbrain.ai/blog/diving-de… Our very own @petetiarks (Rightbrain General Counsel) looks at the legal risks of using the DeepSeek API, comparing their T&Cs with major LLM providers, and exploring alternatives for accessing the open-sourced model, such as self-hosting and cloud platforms offering fine-tuned versions.

Curious about testing another LLM for a task you've got in production? With Rightbrain’s Compare view, you can instantly evaluate models side-by-side. In this demo we run a quick vibe check on DeepSeek R1 vs Claude 3.5 Sonnet on sentiment analysis. This could plug into a pipeline consuming social media data, customer feedback or support channels.

Tweaking LLM prompts? Make changes, save revisions, and test them instantly—see the difference in real time. No more guesswork, just better outputs. 👇 #AI #LLM #ProductManagement #PromptEngineering

📽️ Could your app do image analysis a little better? Do you want to automatically extract structured data from images? Here’s a quick demo on you can build and test an LLM task that generates detailed descriptions from image frames of a 3d virtual tour: > Draft a prompt: Includes specific instructions and the dynamic inputs the task will operate on. > Choose your model: We're using Gemini 2.0 Flash for efficient visual processing. > Define the outputs: Structure responses (like lists and boolean) to match database and UI requirements. > Test your build: Run the task on sample images and see your structured results in action. From here, you can continue iterating further or simply grab the task URL which includes the API call to run the task in a dev environment.

🐳 New week, new model DeepSeek R1 is now available for early access users in the Rightbrain platform. Try it out on your existing LLM tasks or explore completely new use cases. We've been testing it this past week and are impressed by the combo of performance and speed.

With the Rightbrain assistant, you can easily evaluate a task's output against your ideal benchmark. Here's an example: 💡 We used an intelligent troubleshooting task to identify why users might see an error in your app. Integrated seamlessly into user flows, this runs automatically by calling the Rightbrain API. 1️⃣ Start by testing the task on sample data to generate an initial output. 2️⃣ Paste your desired benchmark output to evaluate the results. 3️⃣ The assistant highlights discrepancies and suggests improvements to refine task configuration. 4️⃣ Apply the suggestions and compare the updated prompt's performance. This iterative process ensures your outputs align more closely with your ideal results. 👌

✨ Introducing dynamic inputs, allowing users to customise how certain inputs are handled when performing tasks. Starting with the URL Fetcher, you can now automatically pull web content from a URL and integrate it directly into your LLM tasks, making it easier to work with online data. 🔍 Behind the scenes > system for dynamic input handling. > registry to manage processors. > FastAPI endpoint to explore available options.