sourcebot

s://demo.sourcebot.dev/~/chat/cmdpiuqhu000bpg7s9hprio4w) - “How are shards laid out in memory in the Zoekt code search engine?” (https://demo.sourcebot.dev/~/chat/cmdm9nkck000bod7sqy7c1efb) - "How do I call C from Rust?" (https://demo.sourcebot.dev/~/chat/cmdpjy06g000pnn7ssf4nk60k) You can try it yourself here on our demo site (https://demo.sourcebot.dev/~) or checkout our demo video (https://youtu.be/olc2lyUeB-Q). How is this any different from existing tools like Cursor or Claude code? - Sourcebot solely focuses on code understanding. We believe that, more than ever, the main bottleneck development teams face is not writing code, it’s acquiring the necessary context to make quality changes that are cohesive within the wider codebase. This is true regardless if the author is a human or an LLM. - As opposed to being in your IDE or terminal, Sourcebot is a web app. This allows us to play to the strengths of the web: rich UX and ubiquitous access. We put a ton of work into taking the best parts of IDEs (code navigation, file explorer, syntax highlighting) and packaging them with a custom UX (rich Markdown rendering, inline citations, @ mentions) that is easily shareable between team members. - Sourcebot can maintain an up-to date index of thousands of repos hosted on GitHub, GitLab, Bitbucket, Gerrit, and other hosts. This allows you to ask questions about repositories without checking them out locally. This is especially helpful when ramping up on unfamiliar parts of the codebase or working with systems that are typically spread across multiple repositories, e.g., micro services. - You can BYOK (Bring Your Own API Key) to any supported reasoning model. We currently support 11 different model providers (like Amazon Bedrock and Google Vertex), and plan to add more. - Sourcebot is self-hosted, fair source, and free to use. Under the hood, we expose our existing regular expression search, code navigation, and file reading APIs to a LLM as tool calls. We instruct the LLM via a system prompt to gather the necessary context via these tools to sufficiently answer the users question, and then to provide a concise, structured response. This includes inline citations, which are just structured data that the LLM can embed into it’s response and can then be identified on the client and rendered appropriately. We built this on some amazing libraries like the Vercel AI SDK v5, CodeMirror, react-markdown, and Slate.js, among others. This architecture is intentionally simple. We decided not to introduce any additional techniques like vector embeddings, multi-agent graphs, etc. since we wanted to push the limits of what we could do with what we had on hand. We plan on revisiting our approach as we get user feedback on what works (and what doesn’t). We are really excited about pushing the envelope of code understanding. Give it a try: https://github.com/sourcebot-dev/sourcebot. Cheers!

et me know!

cently went closed source (https://news.ycombinator.com/item?id=41296481). That’s why we built Sourcebot. We designed Sourcebot to be: - Easily deployed: we provide a single, self-contained Docker image (https://github.com/sourcebot-dev/sourcebot/pkgs/container/so...). - Fast & scalable: designed to minimize search times (current average is ~73ms) across many large repositories. - Cross code-host support: we currently support syncing public & private repositories in GitHub and GitLab. - Quality UI: we like to think that a good looking dev-tool is more pleasant to use. - Open source: Sourcebot is free to use by anyone. Under the hood, we use Zoekt (https://github.com/sourcegraph/zoekt) as our code search engine, which was originally authored by Han-Wen Nienhuys and now maintained by Sourcegraph (https://sourcegraph.com/blog/sourcegraph-accepting-zoekt-mai...). Zoekt works by building a trigram index from the source code enabling extremely fast regular expression matching. Russ Cox has a great article on how trigram indexes work if you’re interested: https://swtch.com/~rsc/regexp/regexp4.html In the shorter-term, there are several improvements we want to make, like: - Improving how we communicate indexing progress (this is currently non-existent so it’s not obvious how long things will take) - UX improvements like search history, query syntax highlighting & suggestions, etc. - Small QOL improvements like bookmarking code snippets. - Support for more code hosts (e.g., BitBucket, SourceForge, ADO, etc.) In the longer-term, we want to investigate how we could go beyond just traditional code search by leveraging machine learning to enable experiences like semantic code search (“where is system X located?”) and code explanations (”how does system X interact with system Y?”). You could think of this as a copilot being embedded into Sourcebot. Our hunch is that will be useful to devs, especially when packaged with the traditional code search, but let us know what you think. Give it a try: https://github.com/sourcebot-dev/sourcebot. Cheers!