Your own AI key.
Your engineering context.
Connect Claude, GPT-4o, or Gemini with your own API key — so your AI actually knows what broke and why. Deviera enriches issues, summarises signals, analyses root causes, reviews PRs, and lets you build automations in plain English, all without leaving your workflow.
14-day free trial · No credit card required · BYOK — you pay your provider directly
- Claude Opus 4.7
- Claude Sonnet 4.6
- Claude Haiku 4.5
Best tool-use and function-calling for automation parsing
- GPT-4o
- GPT-4o Mini
- o3-mini
- Gemini 2.0 Flash
- Gemini 1.5 Pro
What the AI layer does
MCP server (Claude integration)
The only engineering intelligence platform with a native MCP server. Claude Desktop and Claude Code can query your signals, list open automations, and create issues directly — without leaving the AI interface.
AI issue enrichment
When Deviera creates a Linear, Jira, or ClickUp issue from a CI failure or stale PR, the AI layer adds a structured description — root cause hypothesis, affected service, suggested fix, and reproduction steps.
Signal summaries
Instead of raw webhook payloads, the Signal Feed shows a one-sentence AI-generated summary of what happened and why it matters — readable by anyone on the team, not just the engineer who owns the repo.
Root cause analysis
When a CI failure or Sentry alert fires, Deviera sends the failure logs, stack trace, and recent commit diff to your LLM and surfaces a root cause hypothesis directly in the issue body.
AI PR comments
On PR open or review request, Deviera posts an AI-generated comment summarising the diff, flagging risky changes, and suggesting test coverage gaps — before a human reviewer picks it up.
Natural-language automation builder
Describe an automation in plain English — "Create a Jira ticket when CI fails on main for more than 30 minutes" — and Deviera's AI parses it into a fully configured trigger → condition → action rule.
How BYOK works
01
Add your API key
Paste your Anthropic, OpenAI, or Gemini key in Settings → AI. Deviera encrypts it at rest with AES-256 and never logs it.
02
Choose your features
Toggle which AI capabilities to enable — issue enrichment, signal summaries, root cause analysis, PR comments. Each can be turned on or off independently.
03
AI works in the background
From the next event onwards, Deviera calls your chosen model. You see richer issues, smarter summaries, and faster automations — with token costs billed directly by your provider.
| Feature | Free | Pro | Team |
|---|---|---|---|
| AI issue enrichment | — | ||
| Signal summaries | — | ||
| Root cause analysis | — | — | |
| AI PR comments | — | — | |
| NL automation builder | — | — | |
| MCP server (Claude) | — |
Frequently asked questions
Does Deviera store my API key?
Your API key is encrypted at rest using AES-256 before storage. Deviera never logs or transmits keys in plaintext.
Which AI providers does Deviera support?
Deviera supports Anthropic (Claude Opus 4.7, Sonnet 4.6, Haiku 4.5), OpenAI (GPT-4o, GPT-4o Mini, o3-mini), and Google Gemini (2.0 Flash, 1.5 Pro). You bring your own API key from any of these providers.
What is BYOK (bring your own key)?
BYOK means you supply your own API key from Anthropic, OpenAI, or Google. Deviera sends requests to the AI provider on your behalf — you control the model, the cost, and the data residency.
What plan do I need for AI features?
AI issue enrichment, signal summaries, and the MCP server are available on Pro and above. Root cause analysis, AI PR comments, and the natural-language automation builder require Team or above.
What is the Deviera MCP server?
The MCP (Model Context Protocol) server exposes Deviera engineering signals as a tool Claude Desktop and Claude Code can call. From inside Claude, you can query open issues, recent CI failures, or create issues without leaving the AI interface.
Connect your provider, pick your features, and let Deviera handle the engineering context — so your AI actually knows what broke and why.