Latest updates to the Sourcegraph Code Intelligence Platform!
In the latest Sourcegraph release for Enterprise Cloud customers we've made improvements to the context fetched from @-mentioned repos, providing greater precision and increasing the likelihood Cody returns higher quality responses.
Single job pods for native Kubernetes Executors have been the default method for the past couple of releases, and based off positive feedback we are now removing multiple job pod configurations as a configuration option.
DeepSeek-V2 was recently introduced as the default autocomplete model for Cody Enterprise customers, and we have implemented optimizations to prompt caching and direct routing that have resulted in improved latency and quality for both single-line and multi-line completions.
Smart Apply now supports the executing of commands in the terminal, allowing users to execute a suggestion Cody provides with the simple click of a button.
OpenAI just unveiled its latest models, 01-preview and o1-mini, which offer enhanced reasoning capabilities for tackling complex tasks. Both models are now available in Cody as a limited release.
We are sunsetting Cody Ignore in this release. Cody Ignore was an experimental feature that was off by default for all users, only enabled via the experimental feature setting in the Cody extension. It lets users specify files or folders in a .cody/ignore file which were then ignored by Cody and disregarded from context.
Cody is built on model interoperability and we aim to provide access to the best and latest models, and today we’re making an update to the default models offered to Enterprise customers: DeepSeek-V2 is the recommended default for autocomplete, while Claude 3.5 Sonnet is now the recommended default for chat and prompts.
The core experience of using Cody is alongside your code in your IDE of choice, but there are often times when you want to interact with Cody via the web. It can be particularly helpful as part of a workflow where you’re performing a search via Code Search and need help from Cody, or if you’re on your mobile device and want to ask it a question. We’re happy to announce that Cody Web is now generally available and includes numerous improvements to make the web experience better and more consistent.
Smart Apply is now available for JetBrains, allowing users to take suggestions from the Cody chat window and near-instantly turning them into diffs in your code.
In our quest for consistency across how you use Cody, JetBrains users now get the same side panel experience found in Cody for VS Code and on the web. Chat lives in the side panel, there are now dedicated tabs for easier navigation, and prompts are now easier to discover and use.
Search Jobs are now generally available, allowing for search queries to be run across your entire codebase where completeness is prioritized over quick response times and results ranking.
Alongside files and repos Cody now lets developers @-mention directories as context, making it easier for users working with larger, more complex repos like monorepos to ensure they're including the best context with their prompts.
Webhooks for Code Monitors are now generally available, alerting users whenever a new search result or change appears via a webhook receiver.
You can now control which parts of a multi-line edit to accept with more granularity. When you ask Cody to edit a block of code, the presented diff will be split into discrete code blocks, and you can accept or reject each diff separately.
You can now control which parts of a multi-line edit to accept with more granularity. When you ask Cody to edit a block of code, the presented diff will be split into discrete code blocks, and you can accept or reject each diff separately.
You can now control which parts of a multi-line edit to accept with more granularity. When you ask Cody to edit a block of code, the presented diff will be split into discrete code blocks, and you can accept or reject each diff separately.
Chat-Oriented Programming (CHOP) allows users to interact with AI to solve problems and write code directly through chat. The new Smart Apply feature enables quick conversion of AI suggestions into code diffs. By pressing Apply, Cody intelligently inserts suggested code directly into code files, streamlining the process from chat to implementation.
Autocomplete has been improved with faster performance and greater accuracy, including a 350-millisecond reduction in latency and a 4% increase in completion acceptance rate. The update features the new DeepSeek-V2 model and is available to Free and Pro users, with full rollout to Enterprise users coming soon.