Skip to main content

About Adrenaline

Indexing

Rather than simply taking your files and stuffing them into an LLM, Adrenaline uses static analysis (e.g. callgraph generation) to build a more sophisticated index of your code.

Graph of knowledge

While looking at a single standalone file, or snippet of code, even if you are unfamiliar with it, you'll be able to figure out what it does. But given a large codebase, with hundreds (or thousands) of dependancy relationships, understanding a single file turns into a months-long process.

In order to understand that graph of dependancies, developers take time onboarding in order to build up a mental map of how the codebase works. This map is a compression of how the codebase actually works.
We take a similar approach with our LLM's - by using the callgraph we build up an index which acts as a mental map for the LLM.

We have made the graphs of knowledge generated for codebase indexes visually explorable.
Come check it out here!

The Adrenaline LLM agents use this index to query and traverse your codebase, in order to generate a more complete understanding of what to include in their articles and diagrams, and what can be left out.

Upgrading your answers

Upgrade the model, and the amount of context used in your answer by turning on Adrenaline Boost (using the Boost toggle). Not only is the model upgraded, but the Adrenaline Agent will take more time searching through your context to build up a refined answer.

Boosted Answers

Beyond getting access to better AI models, with larger context windows, boosing your answer activates a more intelligent Agent.

With Boost on, Adrenaline's agent will use different tools to traverse our index, and build a more complete understanding of the context needed to answer your question.

Generation Speed

Because Adrenaline's Agent is given access to more tools, and takes extra steps to validate it's answer, boosted answers may take longer to generate.

While Boost is deactivated, answers will be generated quickly, on a best-effort basis.

Usage Tiers

Org Repos

Only members subscribed or added to a Teams subscription can add Organization Repos.

FreeLiteMax
Private Repositories1UnlimitedUnlimited
Size Limit5MB25MBUnlimited
Doc Site IndexingExample only5 per-dayUnlimited
Boosted Answers5 per-day25 per-dayUnlimited

Teams and Enterprise

Documentation is almost always out of date, and knowledge is hard to surface. Rather than spending time answering questions in meetings, Adrenaline Teams gives you an AI powered forum to find answers to any question about your org's software.
Every question gets an answer, and every answer will stay up-to-date as your code changes.

Eliminate Time-Consuming Meetings:

Adrenaline Teams gives you a private forum where members can find answers to every question regarding your organization's software. This reduces the need for repetitive explanations in meetings, freeing up valuable time for your team.

Granular Access Control:

Adrenaline Teams provides role-based access control within your organization. This empowers you to tailor information visibility based on individual needs, ensuring a secure and focused knowledge management experience.

Executive summaries

Empower your non-technical team members with clear and concise summaries of complex software functionality. Executive summaries, generated by Adrenaline, bridge the knowledge gap, ensuring everyone understands the core logic driving your software.

This translates to faster decision-making for all stakeholders. By eliminating the need for technical deep dives, Adrenaline streamlines communication and fosters a deeper understanding of how your software delivers business value. Boost team efficiency and drive informed decisions through the power of clear, accessible executive summaries.

Click here to learn more about Teams
https://calendly.com/malik_drabla/30min

Models

We use a mix of GPT-4 and custom finetuned LLMs. To request access to use other models (i.e. Gemini or Claude) on Adrenaline, please contact us at [email protected].
We only offer dedicated hosting of our LLM instances for certain enterprise customers. Schedule a call to learn more.