Red River Insights - The DevTools top10
Red River Insights - June 2024
Dear Friends,
At Red River West, we have a strong tech team and they have allowed us to build one of the most powerful data-driven platform in the VC industry. This month, leveraging RAMP and of course our own developers, we've decided to present you the main trends we see in the DeveloperTools ecosystem.
Dev tools are an essential component of building software, constantly evolving to meet our needs. With the rise of AI, new categories like vector databases and AI-assisted development have emerged, making our work faster and smarter. For instance, 46% of Copilot’s code is now accepted by developers, speeding up coding by 55% (!) (Source). We’ve also noticed a strong trend towards open-source solutions, giving developers more control and flexibility. Startups driving these trends are redefining the future of development with groundbreaking tools and platforms.
Scroll down for our top10 of groundbreaking innovative DevTools startups! (reminder: this is an algorithmic scoring, not a random human judgment!)
RAMP's DevTools Top10
We hope this sparks interesting conversations. If you have any comments or would like to suggest a startup that should be included, feel free to reach out to us. Abel, Chloé, Joseph, Maxime and Olivier will be delighted to discuss these trends and rankings.
(Ranking established on 05/06/2024)
As this top encompasses a lot of different types of startups in the vibrant and dynamic DevTools ecosystem, we've decided to group them by category and we've highlighted the key trends they illustrate:
Databases
They come in many types and are the heart of all applications. But while a majority of the database management systems that are used in production applications today have been created in the early days of the Internet, recently things have been changing.
With the release of OpenAI GPT-3 API and ChatGPT in 2022, we saw a significant surge in interest in Vector Databases. They are a key component of AI-based applications as they allow developers to search and retrieve data tailored to the current context more efficiently. Historical companies from the field such as Pinecone, WeAviate, Qdrant, Chroma, and Vespa have started to build such solutions years in advance and have became cornerstones of early AI-apps, thus receiving substantial funding rounds. Nevertheless, they have recently been seeing a decline in interest from developers.
Indeed, interestingly, historical open-source databases did not get left behind and managed to adapt pretty quickly. The vector storage use case was something that many widely used databases could implement. Early on after the AI craze started, we saw the open-source community implement efficient vector storage in popular databases such as Postgres, and SQLite, … usually through extensions (pgvector, vectorscale, …). This complemented another trend we have been observing recently with companies such as Timescale, ParadeDB, Hydra which are building add-ons and extensions to complement Postgres, turning it into an all-around database.
Another major issue when it comes to databases, aside from data types, is scalability. A lot of open-source companies use that as their business model, as there is no one better to host, maintain, and scale databases than the companies that are developing them, thus allowing for the monetization of open-source projects. This is the case of companies like Qdrant or WeAviate which we mentioned in this ranking.
Some other companies, like Clickhouse, instead decided to off-load their hosting business by having partnerships with hyper scalers (Amazon WebServices, GoogleCloud Platform, Microsoft Azure) and selling through their marketplace. But this is not always the most practical for developers that want to build on top of those databases. Tinybird has solved this by developing a fully serverless architecture on-top of the Clickhouse database management system which allows developers to focus on building applications without having to worry about scaling issues.
Finally, another trend we have observed is the rise of backend-as-a-service offers. This started with Firebase but has recently exploded with Supabase, “the open-source firebase alternative”, and other companies like Appwrite, Nhost, Koyeb… Those companies are not only hosting and managing the database for developers but they also try to build a whole ecosystem around it by offering authentication solutions, file storage through the S3 protocol, and many others, hence simplifying and packaging everything a developer needs to build an application from zero.
Model Management & Observability
Frameworks like Haystack, Langchain, Langfuse, and Jina facilitate the development of large language model (LLM) applications by allowing developers to chain multiple prompts together. However, some within the community feel that these solutions may introduce unnecessary complexity without substantial benefits. Meanwhile, other actors like Mistral and OpenAI are also planning to launch devtools, adding to the uncertainty of who will ultimately succeed in this evolving landscape.
In the realm of observability, Giskard is a notable player in the observability space, particularly for its efforts to combat model bias. It has developed add-ons for LLMs to enhance transparency and fairness in AI models. Giskard's tools help ensure that AI models perform as expected and do not inadvertently reinforce harmful biases.
Model Training
Flower Labs offers a decentralized approach to model training, ensuring that data remains on the device. This approach not only enhances privacy but also reduces the need for large data transfers, making the training process more efficient and secure. Another example is Kestra, an orchestration tool that competes with solutions like Dagster and Prefect. It helps manage complex workflows involved in data processing and machine learning pipelines, ensuring that tasks are executed in the correct sequence and efficiently. We can also mention Hopsworks, specializing in feature preparation and extracting, transforming, loading (ETL) processes. It streamlines the preparation of data for machine learning models, making it easier to create robust and accurate models.
AI-assisted developement tools
AI is increasingly becoming a part of everyday development tools. GitHub Copilot, GitHub Copilot Workspace, Gitstart AI, and Devin are examples of tools integrating AI to enhance coding productivity. Additionally, open-source models like Deepseek and Deepstar coder are gaining attention for their superior performance in coding tasks compared to proprietary solutions.
Devtools Beyond AI
Last but not least, we see other strong companies in DevTools that go beyond AI. A first example is Betterstack, focusing on stack uptime monitoring and log observability. It provides comprehensive tools to ensure that applications remain up and running while offering insights into their performance through detailed logs. Another example is Grafbase, a GraphQL schema aggregator simplifying the management of GraphQL schemas and permissions. It is comparable to Apollo, offering tools that streamline the development and management of GraphQL APIs, making it easier for developers to build robust and secure applications.
More on RAMP's scoring method
The ranking of these startups is based on the estimated momentum of the company but the algorithm does not assess the quality or reliability of the products/solutions developed by these companies!
Find out about the algorithm behind this ranking and the way scores are calculated here: Cheat sheet on RAMP
In case you missed them, our latest top 10s are here:
- June 24
- April 24
- March 24
- February 24
- January 24
All the previous Top 10 are here.
Reverse Lemon Paradox
Olivier shared his thoughts and ideas on the future of secondary transactions in Andre Retterath's Data-Driven Newsletter. You can read the full article here.
Portfolio news
Le Collectionist has launched its brand new app! Go check it here.
After the UK last month, Worldia announced its official launch in the US, and their new partnership with Dream Vacations / CruiseOne, a leading travel agency franchise network in the industry!
Brut. reached 2 billions views in May: this is an all-time record for Brut in terms of views in a single month since their creation in 2016.
The Exploration Company's CEO, Hélène Huby, had the honor to be invited as a member of the delegation accompanying Emmanuel Macron in his trip to Germany. During this trip, celebrating the German French friendship, the signature ceremony for an agreement with Starlab for 3 cargo missions to be provided by The Exploration Company was signed at the French Embassy.
This month, The Exploration Company also signed an agreement with VAST which is entrusting The Exploration Company with one cargo mission.
Robovision's founder, Jonathan Berte, released an article on the impact of AI on the food industry.
You received this newsletter because we met or worked together and we thought you might be interested in an update, if we got it wrong, feel free to unsubscribe with the link at the end!