> For the complete documentation index, see [llms.txt](https://docs.usepylon.com/pylon-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.usepylon.com/pylon-docs/integrations/codebase/devin.md). # Devin Bring Devin into your support workflow. Ask about your codebase, hand off fixes, and track the work, all from Pylon. {% hint style="success" %} Pylon 🤝 Devin partnership offer * Free codebase search, forever. When interfacing with Devin from Pylon, ask-only sessions are free for all users. * $5,000 in Devin credits for existing Devin customers. * $15,000 in Devin credits for new Devin customers who sign up through Pylon. Contact your Pylon or Cognition rep to claim your credits. {% endhint %} Connect Devin by Cognition to Pylon so your team can ask questions about your codebase, or hand work to Devin, without leaving an issue. Cognition usage and billing are not managed by Pylon. ### Overview With the Devin integration, you can: * **Ask Devin about your code.** Find out how a feature works, where a bug might be, or which config is behind a customer's problem, without leaving the issue. * **Hand issues to Devin.** Start a full Devin session to investigate a bug or open a pull request. Follow along and reply to Devin from inside the issue. * **Search your codebase from Copilot and AI Agents.** Codebase search becomes available as a tool for Ask AI and your AI Agents. * **Keep Devin in the loop.** Devin sessions started from Pylon stay linked to their issue. When the customer replies or your team adds new context, Pylon sends it to Devin, so Devin keeps working from the latest information. * **Automate on Devin events.** Run Pylon triggers when Devin opens, merges, or closes a pull request. ### Setup #### 1. Connect Devin 1. In Pylon, go to the **Integrations** and open **Devin**. 2. Click **Connect**, then **Open Devin Settings**. 3. In Devin, connect Pylon from the **Connections** page. Devin authorizes Pylon and finishes the connection for you. When you return to the Devin integration page in Pylon, **OAuth connection** should show **Connected**. {% hint style="info" %} Only Pylon admins can connect or change the Devin integration. {% endhint %} {% hint style="warning" %} Legacy Devin connections relied on API tokens for codebase search, and is not recommended. {% endhint %} #### 2. Add Pylon in Devin Follow the [instructions on the Devin side here](https://docs.devin.ai/integrations/pylon) to configure Pylon in Devin. #### 3. (Optional) Turn on codebase search for Copilot and Agents * **Assist Agent / Copilot:** **Settings → Copilot** * **AI Agents:** **Settings → Agents → Tools** ### Usage #### The Devin applet Each Pylon issue has a Devin applet in the sidebar. It lists every Devin session linked to the issue, whether you started it from Assist Agent or an automation started it. Open a session to: * watch Devin's progress live, * send Devin follow-up messages, * see its final result, or **Open in Devin** to view the full session. There are two kinds of sessions: * **Ask sessions** are read-only. Devin reads your live repositories to answer a question, such as how a feature works, where a bug might be, or which config is behind a customer's problem. * **Full sessions** are for real work: investigating a bug end to end, making changes, or opening a pull request. They run in your Devin workspace like any other Devin session. If Devin can't finish, the session shows **Devin could not finish**, along with the reason. #### Automations Devin automations start sessions on Pylon issues automatically. A common setup is escalating bugs from support to engineering: 1. Pylon flags the issue. A [Pylon Skill](/pylon-docs/agents/skills.md) or AI issue-type detection decides whether a customer report is a real bug with enough detail to act on. If it is, a trigger adds an escalation tag, such as devin-escalate. 2. The tag starts a Devin session. In Devin, create an automation that runs when that tag is added to a Pylon issue. Its instructions define what the session should do, such as reproduce the bug, check logs, and open a fix. 3. Devin investigates. The session reads the issue, tries to reproduce the failure, and opens a pull request where appropriate. Its progress shows up in the issue's Devin applet and as internal notes. 4. An engineer picks it up. The engineer starts with Devin's first pass already done and can continue the session directly. {% hint style="info" %} The tag in your Pylon trigger must exactly match the tag in the Devin automation. To escalate an issue manually, add the tag by hand. {% endhint %} When writing the Skill, separate real bugs from feature requests, how-to questions, and setup problems. When it's unclear, lean toward escalating. The Automations tab on the Devin integration page lists the automations registered with Pylon, including each one's trigger and whether it's enabled, with a link to Open in Devin. You create and manage automations in Devin. #### The @Devin shortcut in Assist Agent {% hint style="info" %} This is for Pylon Agentic customers only. Please reach out to the Pylon team if you are interested in our Agentic features! {% endhint %} In Pylon Assist Agent, quickly consult Devin from the chat by mentioning **@Devin** in your message. Pylon sends Devin the issue context and starts an ask session. The answer shows up in your conversation, and the session appears in the Devin applet. You can move on to another issue while Devin works. From the session card you can open the session, **Open in Devin**, or **Stop** it. When your request needs more than an answer, Assist Agent suggests **Start a full Devin session** and explains why. Accept it to hand the issue to Devin. In read-only mode, Assist Agent can only start ask sessions. Full sessions aren't offered. {% hint style="warning" %} **Full Devin sessions require a Devin seat.** If you don't have a Devin seat, Pylon runs your request as an ask session instead. {% endhint %} #### Devin events and triggers Full sessions stay linked to their issue. Devin posts updates, such as responses and pull requests it opens, merges, or closes, as internal notes in a thread for that session. Pylon tells Devin when the issue gets new activity, such as a customer reply or a new internal note, so it works from the latest context. You can also start Pylon triggers on Devin events: * **Devin pull request opened** * **Devin pull request merged**: for example, reply to the customer when the fix ships * **Devin pull request closed** * **Devin waiting for reply**: for example, notify the assignee so the session isn't left idle ### FAQ **Does @Devin change my code?** Not by default. Ask sessions are read-only. Only a full session, which you approve, can make changes such as opening a pull request. **How is Devin billed?** Full sessions use your Cognition plan, the same as any other Devin session. Ask sessions are complimentary. Pylon doesn't bill for Devin usage. **Why don't I see @Devin in the Assist Agent?** @Devin is only available inside an issue, once Devin is connected. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://docs.usepylon.com/pylon-docs/integrations/codebase/devin.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.