Heym

129 posts

Heym banner
Heym

Heym

@heymrun

Self-hosted, source-available AI workflow automation. https://t.co/TReDbOTJr7

Berlin Katılım Nisan 2026
4 Takip Edilen123 Takipçiler
Sabitlenmiş Tweet
Heym
Heym@heymrun·
Two new Heym templates just dropped: 🔥 1. Slack + Postgres MCP Answer Agent Ask questions in Slack, let an Agent query Postgres through MCP, and post the answer back to the channel. heym.run/templates/slac… Basically text to sql -> read only query -> and return back 2. Daily Telegram Google Analytics MCP Report Run once a day, pull GA metrics through MCP, and send a clean Telegram digest. heym.run/templates/dail… Essentially, good for leads ofc :) MCP powered workflows are getting very real. 🧡
English
0
1
8
284
Heym
Heym@heymrun·
Improved the /chats experience with collapsible tool-call cards, a context usage badge, automatic context compression, and persisted tool-call history across reloads. Cleaner debugging, better visibility into agent actions, and more resilient long-running chats. #AI #Agents #DeveloperTools #UX #OpenSource heym.run
Heym tweet media
English
0
1
6
52
Heym
Heym@heymrun·
New in Heym Traces: full LLM cost observability, built in. Tokens by model, USD spend, latency, error rate, all filterable by time range (1h, 24h, 7d, 30d, all). Per-model pricing syncs from Helicone for 1,100+ models, with per-user overrides and custom model support. #LLMOps #AI #Heym #OpenSource heym.run
Heym tweet media
English
0
0
7
141
Heym
Heym@heymrun·
@TheEcomNomad @ejentum Exactly. The safety-critical branch should be boring and inspectable: pure Python checks panic values first. Critical stops before any sub-agent. Non-critical panels fan out to role-locked agents over the same structured JSON. 🙏🤞
English
0
0
2
19
Aaron ⚡️
Aaron ⚡️@TheEcomNomad·
The panic, value gate staying deterministic is the real move. Most AI health tools flip it backwards, LLM, first on messy stuff, then add guardrails after. You're filtering signal from noise before agents wake up, so parallel interpreters aren't hallucinating differentials from borderline numbers.
English
1
0
2
44
Heym
Heym@heymrun·
Open-source blood panel triage on Heym: 4 cross-lab AI agents by @ejentum Step 1: deterministic 12-marker panic-value gate (pure Python, no LLM). Step 2 (parallel): plain-language interpret, doctor-push, differential. Patient education, not diagnosis. heym.run/templates/bloo… #HealthTech #AIAgents #OpenSource
Heym tweet media
English
3
4
11
123
Heym
Heym@heymrun·
The future of automation is not only trigger → action. AI workflows need agents, RAG, approval checkpoints, traces, evals, and execution control. n8n is strong for general automation. Heym is built for AI-native workflow automation. Different category. Different primitives.
Heym tweet media
English
4
1
9
99
Heym
Heym@heymrun·
With Heym AI Assistant, workflows are not just static diagrams. You can create, edit, and run them from the canvas using natural language. Describe what you need. Let Heym build the flow. Adjust it through chat. Run it directly. Demo: From Prompt to Workflow: Heym AI Assistant Demo youtu.be/5ijD2lGz7zQ?si… via @YouTube
YouTube video
YouTube
English
0
1
8
113
Heym
Heym@heymrun·
Just shipped two new Heym workflow templates for dev and marketing teams. Google Analytics AI Agent -- ask plain-language questions about your GA4 property, get metrics back instantly. heym.run/templates/goog… Google Search Console AI Agent -- rankings, impressions, CTR, indexing status, all from a text box. heym.run/templates/goog… Both run on analytics-mcp and mcp-search-console. Service account setup takes 5 minutes. #MCP #GoogleAnalytics #SearchConsole #AIAgents #SEO #DevTools #Heym #WorkflowAutomation
English
1
0
10
71
Heym
Heym@heymrun·
Generative AI creates an output. Agentic AI drives an outcome. The model can be the same. The difference is the wrapper: planning loop, tools, memory, evaluation, and the ability to act. We wrote a practical guide on when to use each: heym.run/blog/agentic-a… #AgenticAI #GenerativeAI #AIAgents
English
1
0
8
68
Heym
Heym@heymrun·
@briancheong Exactly. MCP solves the model/tool wiring layer. Production is the layer around the call: scoped credentials, auth, rate limits, audit trails, and human approval when needed. In Heym, workflows can use MCP tools or be exposed as MCP tools: heym.run/templates/slac…
Heym tweet media
English
0
0
1
54
Brian Cheong
Brian Cheong@briancheong·
@heymrun The useful part of MCP is standardized tool wiring so agents can swap models without rewriting integrations. The hard part is still auth, rate limits, and auditability.
English
1
0
1
23
Heym
Heym@heymrun·
Most chat 💬 interfaces only answer. Heym Chat can help you control your AI workflow workspace. Provider + model selection Quick Chat actions Natural language requests Document attachments Workflow-aware responses Talk to your automation platform, and let it execute. Demo --> youtube.com/watch?v=Umimz8…
YouTube video
YouTube
English
0
0
6
198