Letter from the Founders

Let’s face it - many aspects of software development are undeniably tedious. Chief among those is diagnosing issues in production systems.


Debugging production errors conjures the same level of enthusiasm as attending your fourth status update meeting this week on a project you barely play a role in. Developing a new feature used by thousands or millions is fun and rewarding. But determining the root cause of an elusive 504 Gateway Timeout error that sporadically affects users? Um, not so much.


We can do better.


Generative AI is writing a new chapter in the ever-evolving story of software. With its ability to make human-like decisions (and even explain its reasoning!) and its uncanny knack for generating new code, new opportunities are emerging for engineering teams to monitor and diagnose their production systems. Imagine a future where automated systems not only detect issues, errors, and anomalies in real-time but also proactively analyze, diagnose, and resolve them. Need to correlate anomalies with historical data patterns to identify root causes? No problem. Need to trace the exact sequence of API calls that led to an error, complete with context from related services? Done. Need to automatically generate and deploy a code fix for a previously unseen edge case? Consider it handled. This is the future of AI-powered error resolution.


With the rise of powerful new language models and the commodification of distributed tracing and machine learning, we see a clear path to realizing our vision—a new intelligent agent that proactively detects issues, analyzes context, and offers informed remediation. CodeComet is our first step towards this vision. Our goal is to eliminate the tedious chores that accompany production software systems, allowing engineering teams to focus on high-leverage tasks and projects. If you catch yourself raising your eyebrows the first time you see our system in action, we’ll do our best to feign surprise. 🙂


To keep it manageable, we’re starting specifically with API production systems. We envision a foundational set of features and functionalities essential to the next generation of API management:


  • Error Resolution: Detect errors in a production API system, like a 5xx or 4xx error, and suggest a code fix to better handle them.


  • Performance Analysis: Detect any performance issue in production, and recommend ways to optimize the system.


  • API Oracle: Ask any question about the behavior or performance of your API in natural language, and receive an answer.


  • Security and Vulnerabilities: Identify security risks and vulnerabilities, and provide actionable remediation steps.


  • Documentation and SDKs: Automatically generate and update API documentation and SDKs.


  • Infrastructure and Resources: Optimize infrastructure usage and resource allocation for your APIs.


  • Mocking: Easily create mock APIs for testing and development purposes.


  • Testing: Create comprehensive test suites to ensure API behavior and reliability.


  • Architecture: Analyze and provide insights into your API architecture for better design and scalability.


  • CI/CD Management: Optimize your CI/CD workflows for faster builds and more reliable deployments.


  • Logging and Monitoring: Enhance logging and monitoring to gain deeper insights into API operations.


We’re sure there’s even more just around the corner. While we have grander ambitions in the long run, we’ve reined them in for our first pass. We won’t launch with everything listed above. As a specific example, we will only support C# and Python to begin with. We want to make sure we know how to walk before we run. So we are building something compact and compelling, that works as advertised, and doesn’t skip those easy-to-neglect edge cases.


Though we are early in this journey, we are committed to building developer-focused tools that bring us closer to the future of software development we see peaking over the horizon. A future where all software systems are performant, reliable, secure, observable…and diagnosable.


If our vision of CodeComet resonates with you, we would love to hear from you.

Rajiv & César

Letter from the Founders

Let’s face it - many aspects of software development are undeniably tedious. Chief among those is diagnosing issues in production systems.


Debugging production errors conjures the same level of enthusiasm as attending your fourth status update meeting this week on a project you barely play a role in. Developing a new feature used by thousands or millions is fun and rewarding. But determining the root cause of an elusive 504 Gateway Timeout error that sporadically affects users? Um, not so much.


We can do better.


Generative AI is writing a new chapter in the ever-evolving story of software. With its ability to make human-like decisions (and even explain its reasoning!) and its uncanny knack for generating new code, new opportunities are emerging for engineering teams to monitor and diagnose their production systems. Imagine a future where automated systems not only detect issues, errors, and anomalies in real-time but also proactively analyze, diagnose, and resolve them. Need to correlate anomalies with historical data patterns to identify root causes? No problem. Need to trace the exact sequence of API calls that led to an error, complete with context from related services? Done. Need to automatically generate and deploy a code fix for a previously unseen edge case? Consider it handled. This is the future of AI-powered error resolution.


With the rise of powerful new language models and the commodification of distributed tracing and machine learning, we see a clear path to realizing our vision—a new intelligent agent that proactively detects issues, analyzes context, and offers informed remediation. CodeComet is our first step towards this vision. Our goal is to eliminate the tedious chores that accompany production software systems, allowing engineering teams to focus on high-leverage tasks and projects. If you catch yourself raising your eyebrows the first time you see our system in action, we’ll do our best to feign surprise. 🙂


To keep it manageable, we’re starting specifically with API production systems. We envision a foundational set of features and functionalities essential to the next generation of API management:


  • Error Resolution: Detect errors in a production API system, like a 5xx or 4xx error, and suggest a code fix to better handle them.


  • Performance Analysis: Detect any performance issue in production, and recommend ways to optimize the system.


  • API Oracle: Ask any question about the behavior or performance of your API in natural language, and receive an answer.


  • Security and Vulnerabilities: Identify security risks and vulnerabilities, and provide actionable remediation steps.


  • Documentation and SDKs: Automatically generate and update API documentation and SDKs.


  • Infrastructure and Resources: Optimize infrastructure usage and resource allocation for your APIs.


  • Mocking: Easily create mock APIs for testing and development purposes.


  • Testing: Create comprehensive test suites to ensure API behavior and reliability.


  • Architecture: Analyze and provide insights into your API architecture for better design and scalability.


  • CI/CD Management: Optimize your CI/CD workflows for faster builds and more reliable deployments.


  • Logging and Monitoring: Enhance logging and monitoring to gain deeper insights into API operations.


We’re sure there’s even more just around the corner. While we have grander ambitions in the long run, we’ve reined them in for our first pass. We won’t launch with everything listed above. As a specific example, we will only support C# and Python to begin with. We want to make sure we know how to walk before we run. So we are building something compact and compelling, that works as advertised, and doesn’t skip those easy-to-neglect edge cases.


Though we are early in this journey, we are committed to building developer-focused tools that bring us closer to the future of software development we see peaking over the horizon. A future where all software systems are performant, reliable, secure, observable…and diagnosable.


If our vision of CodeComet resonates with you, we would love to hear from you.

Rajiv & César

Letter from the Founders

Let’s face it - many aspects of software development are undeniably tedious. Chief among those is diagnosing issues in production systems.


Debugging production errors conjures the same level of enthusiasm as attending your fourth status update meeting this week on a project you barely play a role in. Developing a new feature used by thousands or millions is fun and rewarding. But determining the root cause of an elusive 504 Gateway Timeout error that sporadically affects users? Um, not so much.


We can do better.


Generative AI is writing a new chapter in the ever-evolving story of software. With its ability to make human-like decisions (and even explain its reasoning!) and its uncanny knack for generating new code, new opportunities are emerging for engineering teams to monitor and diagnose their production systems. Imagine a future where automated systems not only detect issues, errors, and anomalies in real-time but also proactively analyze, diagnose, and resolve them. Need to correlate anomalies with historical data patterns to identify root causes? No problem. Need to trace the exact sequence of API calls that led to an error, complete with context from related services? Done. Need to automatically generate and deploy a code fix for a previously unseen edge case? Consider it handled. This is the future of AI-powered error resolution.


With the rise of powerful new language models and the commodification of distributed tracing and machine learning, we see a clear path to realizing our vision—a new intelligent agent that proactively detects issues, analyzes context, and offers informed remediation. CodeComet is our first step towards this vision. Our goal is to eliminate the tedious chores that accompany production software systems, allowing engineering teams to focus on high-leverage tasks and projects. If you catch yourself raising your eyebrows the first time you see our system in action, we’ll do our best to feign surprise. 🙂


To keep it manageable, we’re starting specifically with API production systems. We envision a foundational set of features and functionalities essential to the next generation of API management:


  • Error Resolution: Detect errors in a production API system, like a 5xx or 4xx error, and suggest a code fix to better handle them.


  • Performance Analysis: Detect any performance issue in production, and recommend ways to optimize the system.


  • API Oracle: Ask any question about the behavior or performance of your API in natural language, and receive an answer.


  • Security and Vulnerabilities: Identify security risks and vulnerabilities, and provide actionable remediation steps.


  • Documentation and SDKs: Automatically generate and update API documentation and SDKs.


  • Infrastructure and Resources: Optimize infrastructure usage and resource allocation for your APIs.


  • Mocking: Easily create mock APIs for testing and development purposes.


  • Testing: Create comprehensive test suites to ensure API behavior and reliability.


  • Architecture: Analyze and provide insights into your API architecture for better design and scalability.


  • CI/CD Management: Optimize your CI/CD workflows for faster builds and more reliable deployments.


  • Logging and Monitoring: Enhance logging and monitoring to gain deeper insights into API operations.


We’re sure there’s even more just around the corner. While we have grander ambitions in the long run, we’ve reined them in for our first pass. We won’t launch with everything listed above. As a specific example, we will only support C# and Python to begin with. We want to make sure we know how to walk before we run. So we are building something compact and compelling, that works as advertised, and doesn’t skip those easy-to-neglect edge cases.


Though we are early in this journey, we are committed to building developer-focused tools that bring us closer to the future of software development we see peaking over the horizon. A future where all software systems are performant, reliable, secure, observable…and diagnosable.


If our vision of CodeComet resonates with you, we would love to hear from you.

Rajiv & César