01
Find
Filter the company directory by geography, category, platform, installed apps, and revenue.
Output · target accountsKeep company data, outreach context, and reply state attached to the same record as work moves through the pipeline.
Filter the company directory by geography, category, platform, installed apps, and revenue.
Output · target accountsAdd formulas, HTTP requests, and AI research while keeping every input and result visible.
Output · usable contextBuild email and LinkedIn campaigns from the people and account data you prepared.
Output · live campaignsClassify replies, see the next action, and decide what should be sent or changed.
Output · human decisionNo mystery automation and no context lost between tabs. Each stage shows the data it used, the state it produced, and what needs attention next.
Search the directory before you spend on enrichment. Filter the market by the signals that define your actual customer, then move only selected accounts forward.
Work in a table where each column can reference prior output, call an endpoint, evaluate a formula, or use AI. Inspect the chain instead of trusting a black box.
Write one relevant angle using {{company_news}} and {{tech_stack}}Model selected by workspaceRun email and LinkedIn campaigns, ingest responses, and track the follow-up state. AgentSDR prepares the operational context; your team keeps control of the decision.
38 people · 2 channels
Review the suggested reply. If sent, schedule a follow-up for three business days later.
Configure the database and provider credentials for your environment, inspect the workflow code, and adapt the parts that do not fit the way your team sells.
Explore the repositorygit clone https://github.com/Kandid-ai/AgentSDR_v2.git⌘ CAgentSDR is opinionated about where software should move quickly and where a person should make the call.
AgentSDR is an operating product, not a magic button. Here is the practical version.
It brings account discovery, enrichment tables, email and LinkedIn campaign workflows, and reply operations into one application. Your team may still keep a separate system of record, but less work needs to move through spreadsheets and point tools.
Founder-led sales teams, small GTM teams, and agencies that run outbound themselves. It is especially useful when the team wants to inspect, host, and adapt the workflow rather than accept a fixed SaaS process.
Yes. The workspace supports bring-your-own-key model configuration, so the deployment controls the provider credentials and model selection available to operators.
AgentSDR runs as a Next.js application backed by Postgres. Configure the required environment variables and the external services you intend to use, then run it on the infrastructure your team manages.
No. AgentSDR automates operational work such as scheduling, ingestion, and classification while preserving review points for drafts, classifications, and sensitive follow-up actions.
Both channels are supported, with channel-specific campaign and message surfaces. The shared people and CRM context connects the resulting records and next actions.
Connect the services you already use. Keep the data and decisions in a system you understand.
git clone https://github.com/Kandid-ai/AgentSDR_v2.git