Job application tracker and AI-powered job search assistant. Helps job seekers manage their search journey with AI resume review, job matching, task logging, and application analytics, all while keeping your data private.
TypeScript ai-agentai-resume-analyzerapplication-trackerapplication-tracking-systematscareerjob-applicationjob-automationjob-huntingjob-searchjob-seeker-toolsjob-trackerjob-trackingjobsearchjobseekersmcpollamaresumeresume-reviewself-hosted
import type { DescriptionCompleteness } from "@/models/job.model";
import type { JobMatchScores, ResumeScores } from "@/models/ai.schemas";
// Dependency-free by design: client components import these types, so nothing
// here may pull in Prisma or server-only code. Type-only imports are erased.
export type PageContext = {
route?: string;
jobId?: string;
resumeId?: string;
};
// Custom data part, not a FileUIPart: a file part is a standard type that
// convertToModelMessages would carry to the model in full, defeating head
// truncation. Data parts are dropped, which is why the head is injected.
export const AGENT_PASTE_PART_TYPE = "data-paste";
export type AgentPastePartData = {
id: string;
text: string;
chars: number;
truncated: boolean;
consumed?: boolean;
};
export type AgentPastePart = {
type: typeof AGENT_PASTE_PART_TYPE;
id?: string;
data: AgentPastePartData;
};
export function isAgentPastePart(part: unknown): part is AgentPastePart {
const candidate = part as AgentPastePart | undefined;
return (
!!candidate &&
candidate.type === AGENT_PASTE_PART_TYPE &&
typeof candidate.data?.text === "string"
);
}
export type AgentResolvedEntity = {
id: string;
label: string;
created: boolean;
};
// What add_job returns to the model AND to the result card. The card composes
// its own copy from these fields — never from createJobFromNames' message,
// which is MCP-facing protocol text naming tools the chat does not expose.
export type AgentAddJobResult = {
created: boolean;
jobId?: string;
duplicateOf?: { id: string; title: string; company: string };
resolutions: AgentResolvedEntity[];
descriptionSource?: "pasted" | "model";
descriptionChars?: number;
descriptionCompleteness?: DescriptionCompleteness;
validationError?: string;
// The card renders validationError verbatim, so a recovery instruction
// written at the model reaches the user as prose about itself. Set this
// wherever the two audiences need different words; Zod's own messages name
// the field and its values and read fine unaided, so they leave it unset.
displayError?: string;
};
// Which rule picked the resume. Surfaced so the result card can say "your
// default resume" instead of leaving the user guessing which one was read.
export type AgentResumeSource = "named" | "page" | "default" | "only";
// What get_resume returns to the model AND to the result card. resumeText is
// the only large field; the card must never render it.
export type AgentGetResumeResult =
| {
status: "ok";
resumeId: string;
title: string;
resumeText: string;
chars: number;
truncated: boolean;
source: AgentResumeSource;
ambiguousTitle?: boolean;
}
| { status: "needs_selection"; resumes: { id: string; title: string }[] }
| { status: "no_resumes" }
| { status: "unreadable"; title: string; reason: string };
// Transient stream part carrying a nested generation's tokens. It is never
// persisted: the finished artifact lands in the tool result, and keeping both
// would duplicate it in storage and in model context. Shared by review_resume
// and match_job — the toolCallId key is what keeps two of them apart.
export const AGENT_NESTED_STREAM_PART_TYPE = "data-nested-stream";
export type AgentNestedStreamData = { delta: string };
// What review_resume returns to the model AND to the result card. The resume
// text is deliberately absent — follow-ups answer from `body`, not from the
// serialization.
export type AgentReviewResumeResult =
| {
status: "ok";
resumeId: string;
title: string;
scores: ResumeScores;
body: string;
saved: boolean;
saveError?: string;
}
| { status: "needs_selection"; resumes: { id: string; title: string }[] }
| { status: "no_resumes" }
| { status: "unreadable"; title: string; reason: string }
| { status: "generation_failed"; title: string; reason: string };
// What match_job returns to the model AND to the result card. Neither the
// resume text nor the job description is here — follow-ups answer from `body`.
export type AgentMatchJobResult =
| {
status: "ok";
jobId: string;
jobTitle: string;
company: string;
resumeId: string;
resumeTitle: string;
scores: JobMatchScores;
body: string;
saved: boolean;
saveError?: string;
}
| { status: "no_job" }
| { status: "needs_selection"; resumes: { id: string; title: string }[] }
| { status: "no_resumes" }
| { status: "unreadable"; what: "job" | "resume"; title: string; reason: string }
| { status: "generation_failed"; jobTitle: string; reason: string };
// What generate_cover_letter returns to the model AND to the result card. The
// letter body is here so a follow-up ("make it shorter") answers from it; the
// resume text and the job description are not.
export type AgentCoverLetterResult =
| {
status: "ok";
jobId: string;
jobTitle: string;
company: string;
resumeId: string;
resumeTitle: string;
body: string;
coverLetterId?: string;
coverLetterTitle?: string;
saved: boolean;
saveError?: string;
}
| { status: "no_job" }
| { status: "no_description"; jobTitle: string }
| { status: "needs_selection"; resumes: { id: string; title: string }[] }
| { status: "no_resumes" }
| { status: "unreadable"; what: "job" | "resume"; title: string; reason: string }
| { status: "generation_failed"; jobTitle: string; reason: string };
// Tools that run their own generation and stream it as transient parts. The
// transcript and the provider both branch on this rather than on a chain of
// name comparisons that grows with every nested tool.
export const AGENT_NESTED_TOOLS = [
"review_resume",
"match_job",
"generate_cover_letter",
] as const;
export function isNestedTool(name: string): boolean {
return (AGENT_NESTED_TOOLS as readonly string[]).includes(name);
}
// Tools that end the turn. The result card renders deterministically from
// structured fields, so a second generation just to narrate it is 10-30s of
// local inference for a sentence that could be wrong. A tool that runs its
// own generation MUST be listed here or a single turn can chain two of them
// and blow the turn timeout, discarding both.
export const AGENT_CHAT_TERMINAL_TOOLS = [
"add_job",
...AGENT_NESTED_TOOLS,
] as const;
// add_job is terminal once it has SETTLED — written, hit a duplicate, or
// asked for approval — not merely once it has been called. A call that cannot
// write skips approval and so fails inside the same step that made it, and
// hasToolCall would end the turn on the one step the model needs to retry in,
// leaving the user a red card and no answer.
export function addJobSettled(step: {
toolCalls?: readonly { toolName: string; invalid?: boolean }[];
toolResults?: readonly { toolName: string; output?: unknown }[];
}): boolean {
const call = step.toolCalls?.find((c) => c.toolName === "add_job");
if (!call) return false;
// A call whose input failed the tool's own schema never reaches execute:
// the SDK marks it invalid and emits a tool-error, so it produces no
// tool-result and would otherwise read as a pending approval below. It is
// the same retryable failure as a validationError, one layer up.
if (call.invalid) return false;
const result = step.toolResults?.find((r) => r.toolName === "add_job");
// No result yet means the call is waiting on the approval card, and nothing
// more can happen this turn.
if (!result) return true;
return !(result.output as AgentAddJobResult | undefined)?.validationError;
}