Ashley Budgie
AI Agent Engineer | Voice agents, automations & integrations
New to Contra
Ashley is ready for their next project!
AI Hotkey Assistant: Desktop AI Tool
Copy-pasting to ChatGPT dozens of times a day was killing my focus: select text, switch tabs, paste, wait, copy the result, switch back, paste again. I built a Windows desktop tool that collapses the whole loop into one hotkey.
Select any text anywhere, press a hotkey, and it is transformed in place by the AI service of your choice, using your own prompt files: fix spelling, rewrite, translate, summarize, or anything you can describe in a prompt. A second hotkey path reads any selected text aloud through text-to-speech, with keyboard control of the audio.
The app lives in the system tray with a full settings interface: multiple AI services (any OpenAI-compatible API, plus ElevenLabs for voice), per-action hotkeys, prompts as editable files, and model lists fetched from each provider. API keys are stored in the Windows credential manager, never in a config file.
Built for my own workflow and shipped as v2.0.0.
Stack: Python, pywin32 (global hotkeys + tray), customtkinter (settings UI), keyring (credentials), OpenAI-compatible chat APIs, ElevenLabs TTS.
0
15
Inbound AI Voice Agent, Configured in GoHighLevel (Plus a Custom Agent in Our Own Dialer)
I built an inbound AI voice agent in GHL that answers the client's number. It greets the caller, works the conversation from a script with built-in objection handling, and books or routes the call based on how it goes. Every call writes back to the CRM: contact record, opportunity stage, timeline note. The team sees exactly what happened on every call without touching a thing. No new software to learn, it's configured in the CRM they already run on.
And when a client needs the outbound side, I've built that too. A custom AI voice agent from the pipeline up. Pipecat with OpenAI Realtime, speech to text, LLM, text to speech, all in a single WebSocket, embedded right into our own power dialer. A sales team dials prospects in parallel and hands the conversation to the AI: script, objection handling, voicemail branch, and the outcome written straight back to the CRM. The whole pipeline is covered by a 91-case automated test suite, and if the AI service goes down, the call degrades gracefully instead of crashing.
Inbound on GHL, outbound in our dialer. Same voice AI capability, in the shape the client's workflow needs.
0
35
Compliance-Aware Email + SMS Follow-Up Automation.
A sales team was making cold calls and losing the follow-up. Leads that asked to be contacted later had no reliable path back into the funnel, and the team had no way to guarantee they were respecting opt-in rules. I built a follow-up automation that handles the timing and the compliance in one system.
The workflow triggers when a call ends with a follow-up disposition, then runs a 3-day timed sequence of email and SMS touches. Before any message goes out, the system checks the contact's opt-in status, and any contact on the blacklist is skipped entirely. Follow-ups only ever go to people who asked for them.
Every step is visible in the CRM timeline, so the team can see exactly where each lead is in the sequence, and a human can take over at any point without breaking the flow. It turns "we'll follow up" from a promise into a guarantee, without adding another manual task to the day.
1
71
I build custom Power Dialer with Parallel Calling and CRM Integration
0
60